input
stringlengths 28
631
| output
stringlengths 67
772
| instruction
stringclasses 1
value |
---|---|---|
In 2003, which team broke Australia's winning streak of the World Netball Tournaments since 1971? | Plan: Search for more information about the World Netball Tournament
#E1 = Wikipedia[World Netball Tournament]
Plan: Search for information about Australia's winning streak in the tournament
#E2 = Wikipedia[Australia's winning streak in the World Netball Tournament]
Plan: Find out the name of the team that broke Australia's winning streak in 2003
#E3 = LLM[What team broke Australia's winning streak in the World Netball Tournament in 2003? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which political group imposed dictatoriship in France after the revolution? | Plan: Search for more information about the French Revolution
#E1 = Wikipedia[French Revolution]
Plan: Search for more information about political groups in France after the revolution
#E2 = Wikipedia[Political Groups in France after the Revolution]
Plan: Find out the name of the political group that imposed dictatorship in France after the revolution.
#E3 = LLM[What is the name of the political group that imposed dictatorship in France after the revolution? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What Big 12 Conference basketball team did Myles Turner play for?
| Plan: Search for more information about Myles Turner.
#E1 = Wikipedia[Myles Turner]
Plan: Find out which Big 12 Conference basketball team Myles Turner played for.
#E2 = LLM[Which Big 12 Conference basketball team did Myles Turner play for? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What was the original stage name of the artist who recorded the song element ?
| Plan: Search for more information about the song element.
#E1 = Wikipedia[Element (song)]
Plan: Find out the artist who recorded the song element.
#E2 = LLM[Who recorded the song element? Given context: #E1]
Plan: Find out the original stage name of the artist.
#E3 = LLM[What was the original stage name of #E2? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which star of "Live Free or Die" was born on March 20th, 1970? | Plan: Search for more information about the movie "Live Free or Die".
#E1 = Wikipedia["Live Free or Die"]
Plan: Find out which actor starred in the movie.
#E2 = LLM[Which actor starred in the movie "Live Free or Die"? Given context: #E1]
Plan: Search for more information about the actor.
#E3 = Wikipedia[#E2]
Plan: Find out the actor's date of birth.
#E4 = LLM[What is the date of birth of #E2? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Are Cho Kyuhyun and Layne Staley both singers? | Plan: Search for more information about Cho Kyuhyun.
#E1 = Wikipedia[Cho Kyuhyun]
Plan: Search for more information about Layne Staley.
#E2 = Wikipedia[Layne Staley]
Plan: Compare the two and determine if they are both singers.
#E3 = LLM[ | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The origins of which country can be traced to the forming of a defensive league by Uri, Schwyz and Lower Unterwalden in 1521? | Plan: Search for more information about Uri, Schwyz, and Lower Unterwalden
#E1 = Wikipedia[Uri, Schwyz, Lower Unterwalden]
Plan: Search for more information about the defensive league formed by the three
#E2 = Wikipedia[defensive league formed by Uri, Schwyz and Lower Unterwalden]
Plan: Find out the name of the country which traces its origins to the defensive league formed by the three
#E3 = LLM[What is the name of the country which traces its origins to the defensive league formed by Uri, Schwyz and Lower Unterwalden? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Rolfe R. Kent who produced the original score of Legally Blonde was born in what country?
| Plan: Search for more information about Rolfe R. Kent.
#E1 = Wikipedia[Rolfe R. Kent]
Plan: Find out what country Rolfe R. Kent was born in.
#E2 = LLM[What country was Rolfe R. Kent born in? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which playwright was the designer of Blenheim Palace and Castle Howard ? | Plan: Search for more information about Blenheim Palace
#E1 = Wikipedia[Blenheim Palace]
Plan: Search for more information about Castle Howard
#E2 = Wikipedia[Castle Howard]
Plan: Find out the name of the playwright who designed both Blenheim Palace and Castle Howard.
#E3 = LLM[Who is the playwright who designed Blenheim Palace and Castle Howard? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In 2012 who became the oldest winner of an Oscar, for his role in 'Beginners'? | Plan: Search for more information about the movie 'Beginners'
#E1 = Wikipedia[Beginners]
Plan: Search for more information about the oldest Oscar winner
#E2 = Wikipedia[Oldest Oscar winner] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Kathy Jordan and Jack Sock, share which occupation?
| Plan: Search for more information about Kathy Jordan and Jack Sock.
#E1 = Wikipedia[Kathy Jordan], Wikipedia[Jack Sock]
Plan: Find out what occupation Kathy Jordan and Jack Sock share.
#E2 = LLM[What occupation do Kathy Jordan and Jack Sock share? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Did the Pharaoh Hound and German Pinscher both originate from the same country?
| Plan: Search for more information about Pharaoh Hound.
#E1 = Wikipedia[Pharaoh Hound]
Plan: Search for more information about German Pinscher.
#E2 = Wikipedia[German Pinscher]
Plan: Find out the country of origin for both Pharaoh Hound and German Pinscher.
#E3 = LLM[What is the country of origin for Pharaoh Hound and German Pinscher? Given context: #E1, #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Are The Mowgli's and Bodyjar both alternative rock bands? | Plan: Search for more information about The Mowgli's.
#E1 = Wikipedia[The Mowgli's]
Plan: Search for more information about Bodyjar.
#E2 = Wikipedia[Bodyjar]
Plan: Compare the two and determine if they are both alternative rock bands.
#E3 = LLM[ | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which US First Lady had an English Springer Spaniel called Millie? | Plan: Search for more information about US First Ladies
#E1 = Wikipedia[US First Ladies]
Plan: Search for more information about Millie
#E2 = Wikipedia[Millie (dog)]
Plan: Find out which US First Lady had an English Springer Spaniel called Millie.
#E3 = LLM[Which US First Lady had an English Springer Spaniel called Millie? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What does the retail chain founded by Lars Larsen sell?
| Plan: Search for more information about Lars Larsen.
#E1 = Wikipedia[Lars Larsen]
Plan: Find out what the retail chain founded by Lars Larsen sells.
#E2 = LLM[What does the retail chain founded by Lars Larsen sell? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The person William Mein Smith worked as the Surveyor General for from 1840 to 1843 was born on what date?
| Plan: Search for more information about William Mein Smith.
#E1 = Wikipedia[William Mein Smith]
Plan: Find out the date of birth of William Mein Smith.
#E2 = LLM[What is the date of birth of William Mein Smith? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
A Soviet film adaptation of which Shakespearean play, directed by Grigori Kozintsev, with black-and-white photography, a score by Shostakovich and a script based on a translation by Boris Pasternak, was made by Mosfilm in 1971? | Plan: Search for more information about Grigori Kozintsev
#E1 = Wikipedia[Grigori Kozintsev]
Plan: Search for more information about films directed by Grigori Kozintsev
#E2 = Wikipedia[Films directed by Grigori Kozintsev]
Plan: Find out the name of the Shakespearean play that was adapted into a Soviet film by Mosfilm in 1971
#E3 = LLM[What is the name of the Shakespearean play that was adapted into a Soviet film by Mosfilm in 1971? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Rear Admiral (UH) Raymond A. Spicer, United States Navy, is the former director of the White House Military Office and reported as Deputy Assistant to the President and Director, White House Military Office in March 2007, and was succeeded by which American politician and member of the Democratic Party who served as the Director of the White House Military Office from January to May 2009, as the 17th United States Secretary of the Army? | Plan: Search for more information about Rear Admiral (UH) Raymond A. Spicer.
#E1 = Wikipedia[Rear Admiral (UH) Raymond A. Spicer]
Plan: Find out who succeeded Raymond A. Spicer as the Director of the White House Military Office.
#E2 = LLM[Who succeeded Raymond A. Spicer as the Director of the White House Military Office? Given context: #E1]
Plan: Search for more information about the person who succeeded Raymond A. Spicer.
#E3 = Wikipedia[#E2]
Plan: Find out which American politician and member of the Democratic Party served as the 17th United States Secretary of the Army.
#E4 = LLM[Which American politician and member of the Democratic Party served as the 17th United States Secretary of the Army? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"Stand by Me" is a song by the English rock band Oasis from an album released on which day ?
| Plan: Search for more information about the song "Stand by Me".
#E1 = Wikipedia["Stand by Me" (Oasis song)]
Plan: Find out the day the album was released.
#E2 = LLM[What day was the album released? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which Italian opera has a title meaning 'The strolling players (or clowns)'? | Plan: Search for more information about Italian operas
#E1 = Wikipedia[Italian opera]
Plan: Search for more information about the meaning of 'The strolling players (or clowns)'
#E2 = Wikipedia['The strolling players (or clowns)']
Plan: Find out the name of the Italian opera with a title meaning 'The strolling players (or clowns)'.
#E3 = LLM[What is the name of the Italian opera with a title meaning 'The strolling players (or clowns)'? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
When did the person that rebuilt the Temple of Minerva Chalcidica rule Rome?
| Plan: Search for more information about the Temple of Minerva Chalcidica.
#E1 = Wikipedia[Temple of Minerva Chalcidica]
Plan: Find out who rebuilt the Temple of Minerva Chalcidica.
#E2 = LLM[Who rebuilt the Temple of Minerva Chalcidica? Given context: #E1]
Plan: Find out when the person who rebuilt the Temple of Minerva Chalcidica ruled Rome.
#E3 = LLM[When did #E2 rule Rome? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which retail guru presented the TV series Mary Queen of Shops ? | Plan: Search for more information about Mary Queen of Shops
#E1 = Wikipedia[Mary Queen of Shops]
Plan: Search for the presenter of Mary Queen of Shops
#E2 = Wikipedia[Mary Queen of Shops presenter]
Plan: Find out the name of the retail guru who presented Mary Queen of Shops.
#E3 = LLM[What is the name of the retail guru who presented Mary Queen of Shops? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Crouch End is a horror story by Stephen King, originally published in "New Tales of the Cthulhu Mythos" (1980), it contains distinct references to the horror fiction of which American author who achieved posthumous fame through his influential works of horror fiction?
| Plan: Search for more information about Crouch End.
#E1 = Wikipedia[Crouch End]
Plan: Find out which American author Stephen King referenced in Crouch End.
#E2 = LLM[Which American author did Stephen King reference in Crouch End? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"In which film did Clint Eastwood sing the song ""I Talk To The Trees'?" | Plan: Search for more information about Clint Eastwood
#E1 = Wikipedia[Clint Eastwood]
Plan: Search for more information about the song ""I Talk To The Trees'
#E2 = Wikipedia["I Talk To The Trees"]
Plan: Find out the name of the film in which Clint Eastwood sang the song ""I Talk To The Trees'
#E3 = LLM[What is the name of the film in which Clint Eastwood sang the song ""I Talk To The Trees'? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Who was the first Irishman to win the Tour de France? | Plan: Search for more information about Tour de France
#E1 = Wikipedia[Tour de France]
Plan: Search for Irish cyclists who have won Tour de France
#E2 = Wikipedia[Irish cyclists who have won Tour de France]
Plan: Find out the name of the first Irishman to win the Tour de France
#E3 = LLM[What is the name of the first Irishman to win the Tour de France? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What country would you have to visit if you wanted to 'Row on the Po'? | Plan: Search for more information about the Po River
#E1 = Wikipedia[Po River]
Plan: Find out what country the Po River is located in
#E2 = LLM[What country is the Po River located in? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What was the population according to the 2010 census of the census-designated place where Max Loughan created a crystal radio? | Plan: Search for more information about Max Loughan.
#E1 = Wikipedia[Max Loughan]
Plan: Find out the census-designated place where Max Loughan created a crystal radio.
#E2 = LLM[What is the census-designated place where Max Loughan created a crystal radio? Given context: #E1]
Plan: Search for more information about the census-designated place.
#E3 = Wikipedia[#E2]
Plan: Find out the population according to the 2010 census of the census-designated place.
#E4 = LLM[What was the population according to the 2010 census of the census-designated place #E2? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The Mohs scale measures what? | Plan: Search for more information about Mohs scale
#E1 = Wikipedia[Mohs scale] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In which period the father of Queen Anne-Marie reign Denmark? | Plan: Search for more information about Queen Anne-Marie.
#E1 = Wikipedia[Queen Anne-Marie]
Plan: Find out the name of Queen Anne-Marie's father.
#E2 = LLM[What is the name of Queen Anne-Marie's father? Given context: #E1]
Plan: Search for more information about the father of Queen Anne-Marie.
#E3 = Wikipedia[#E2]
Plan: Find out in which period the father of Queen Anne-Marie reign Denmark.
#E4 = LLM[In which period did the father of Queen Anne-Marie reign Denmark? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Noble is a 2014 film written and directed by Stephen Bradley about the true life story of Christina Noble, a children's rights campaigner, charity worker and writer, it stars Sarah Greene, an Irish actress and singer, best known for portraying Helen McCormick in the West End and Broadway productions of what? | Plan: Search for more information about Christina Noble.
#E1 = Wikipedia[Christina Noble]
Plan: Find out what movie Stephen Bradley wrote and directed about Christina Noble.
#E2 = LLM[What movie did Stephen Bradley write and direct about Christina Noble? Given context: #E1]
Plan: Search for more information about the movie.
#E3 = Wikipedia[#E2]
Plan: Find out who Sarah Greene portrays in the movie.
#E4 = LLM[Who does Sarah Greene portray in the movie #E2? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which island group in the Indian Ocean, a popular tourist destination, has the city of Victoria as its capital ? | Plan: Search for more information about the Indian Ocean
#E1 = Wikipedia[Indian Ocean]
Plan: Search for more information about popular tourist destinations in the Indian Ocean
#E2 = Wikipedia[Popular tourist destinations in the Indian Ocean]
Plan: Find out the name of the island group with the city of Victoria as its capital
#E3 = LLM[What is the name of the island group with the city of Victoria as its capital? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What musical-sitcom starring Shirley Jones did Bill Manhoff write for?
| Plan: Search for more information about Bill Manhoff.
#E1 = Wikipedia[Bill Manhoff]
Plan: Find out what musical-sitcom Bill Manhoff wrote for.
#E2 = LLM[What musical-sitcom did Bill Manhoff write for? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Clarence "Keith" Marquis Mitchell, is a former American football linebacker in which rganization,who played for the Houston Texans, a professional American football team based in Houston, Texas? | Plan: Search for more information about Clarence "Keith" Marquis Mitchell.
#E1 = Wikipedia[Clarence "Keith" Marquis Mitchell]
Plan: Find out which organization he played for.
#E2 = LLM[Which organization did Clarence "Keith" Marquis Mitchell play for? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which famous singer was once married to Renate Blauel? | Plan: Search for more information about Renate Blauel
#E1 = Wikipedia[Renate Blauel]
Plan: Find out which famous singer was once married to Renate Blauel.
#E2 = LLM[Which famous singer was once married to Renate Blauel? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"What is the Spanish word for ""house""?" | Plan: Search for the Spanish word for "house"
#E1 = Wikipedia["Spanish language" and "house"] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Late actress Greta Garbo was born in which European country? | Plan: Search for more information about Greta Garbo
#E1 = Wikipedia[Greta Garbo]
Plan: Find out the European country in which Greta Garbo was born
#E2 = LLM[In which European country was Greta Garbo born? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What american actor, who starred in the film What They Had, also received a Tony Award nomination for Best Featured Actor in a Play in 2016?
| Plan: Search for more information about What They Had.
#E1 = Wikipedia[What They Had]
Plan: Find out the actor who starred in the film.
#E2 = LLM[Which actor starred in the film What They Had? Given context: #E1]
Plan: Find out if the actor received a Tony Award nomination in 2016.
#E3 = LLM[Did #E2 receive a Tony Award nomination for Best Featured Actor in a Play in 2016? Given context: #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What human organ makes four new words when prefixing Head, Diver, Flint and Tight? | Plan: Find out the definition of each word
#E1 = Wikipedia[Head], Wikipedia[Diver], Wikipedia[Flint], Wikipedia[Tight]
Plan: Identify the common element in the words
#E2 = LLM[What is the common element in the words Head, Diver, Flint and Tight? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What type of entertainment does Four Star Playhouse and David Niven have in common?
| Plan: Search for more information about Four Star Playhouse.
#E1 = Wikipedia[Four Star Playhouse]
Plan: Find out what type of entertainment Four Star Playhouse provides.
#E2 = LLM[What type of entertainment does Four Star Playhouse provide? Given context: #E1]
Plan: Search for more information about David Niven.
#E3 = Wikipedia[David Niven]
Plan: Find out what type of entertainment David Niven is associated with.
#E4 = LLM[What type of entertainment is David Niven associated with? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The jazz classic Mood Indigo is particularly associated with whose orchestra? | Plan: Search for more information about Mood Indigo
#E1 = Wikipedia[Mood Indigo]
Plan: Search for more information about jazz orchestras
#E2 = Wikipedia[Jazz Orchestras]
Plan: Find out which orchestra is particularly associated with Mood Indigo
#E3 = LLM[Which orchestra is particularly associated with Mood Indigo? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The Kings of Summer stars which actor who played Ron Swanson on "Parks and Recreation"? | Plan: Search for more information about The Kings of Summer.
#E1 = Wikipedia[The Kings of Summer]
Plan: Find out which actor starred in The Kings of Summer.
#E2 = LLM[Which actor starred in The Kings of Summer? Given context: #E1]
Plan: Find out which TV show the actor starred in.
#E3 = LLM[Which TV show did the actor who starred in The Kings of Summer star in? Given context: #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Permanent software programmed into a read-only memory is generally called what ' ...ware'? | Plan: Search for more information about permanent software programmed into read-only memory
#E1 = Wikipedia[Permanent software programmed into read-only memory]
Plan: Find out the term used to describe this type of software.
#E2 = LLM[What is the term used to describe permanent software programmed into a read-only memory? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which ancient philosopher's works are collected in 'The Analects'? | Plan: Search for more information about The Analects
#E1 = Wikipedia[The Analects]
Plan: Identify the philosopher who wrote The Analects
#E2 = LLM[Who wrote The Analects? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
AutoTrader.co.za is a site for the company that was once a subsidiary of the company that was founded by whom? | Plan: Search for more information about AutoTrader.co.za.
#E1 = Wikipedia[AutoTrader.co.za]
Plan: Find out which company AutoTrader.co.za was once a subsidiary of.
#E2 = LLM[Which company was once a subsidiary of the company that AutoTrader.co.za was part of? Given context: #E1]
Plan: Find out which company the subsidiary was founded by.
#E3 = LLM[Who founded the company that AutoTrader.co.za was once a subsidiary of? Given context: #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Nigeria's 14th and current (at 2014) presidential head of state is (What optimistic expression?) Jonathan? | Plan: Search for more information about Nigeria's current president
#E1 = Wikipedia[Nigeria current president]
Plan: Search for more information about Jonathan
#E2 = Wikipedia[Jonathan]
Plan: Find out the optimistic expression associated with Jonathan
#E3 = LLM[What optimistic expression is associated with Jonathan? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which is located in Hengshui, Bei'an or Shenzhou City? | Plan: Search for more information about Hengshui.
#E1 = Wikipedia[Hengshui]
Plan: Search for more information about Bei'an.
#E2 = Wikipedia[Bei'an]
Plan: Search for more information about Shenzhou City.
#E3 = Wikipedia[Shenzhou City]
Plan: Compare the three and determine which is located in Hengshui, Bei'an or Shenzhou City.
#E4 = LLM[ | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which major Russian composer of the 20th century wrote a piece that was premiered by the Beethoven Quartet in 1964?
| Plan: Search for more information about the Beethoven Quartet.
#E1 = Wikipedia[Beethoven Quartet]
Plan: Find out which major Russian composer of the 20th century wrote a piece that was premiered by the Beethoven Quartet in 1964.
#E2 = LLM[Which major Russian composer of the 20th century wrote a piece that was premiered by the Beethoven Quartet in 1964? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which former 'Liverbird' went on to play The District Nurse' on T.V? | Plan: Search for more information about Liverbird
#E1 = Wikipedia[Liverbird]
Plan: Search for more information about The District Nurse
#E2 = Wikipedia[The District Nurse]
Plan: Ask the LLM to identify the former Liverbird who went on to play The District Nurse
#E3 = LLM[Who was the former Liverbird who went on to play The District Nurse? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Born on 20th July 1956, who was the original drummer in the punk rock band The Sex Pistols? | Plan: Search for more information about The Sex Pistols
#E1 = Wikipedia[The Sex Pistols]
Plan: Search for more information about the original drummer
#E2 = Wikipedia[original drummer of The Sex Pistols] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What is the capital of Paraguay | Plan: Search for more information about Paraguay
#E1 = Wikipedia[Paraguay]
Plan: Find out the name of the capital of Paraguay
#E2 = LLM[What is the capital of Paraguay? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In Greek Cuisine which dish consists of vine leaves stuffed with rice and vegetables? | Plan: Search for more information about Greek Cuisine
#E1 = Wikipedia[Greek Cuisine]
Plan: Search for more information about dishes in Greek Cuisine
#E2 = Wikipedia[Dishes in Greek Cuisine]
Plan: Find out the name of the dish consisting of vine leaves stuffed with rice and vegetables.
#E3 = LLM[What is the name of the dish consisting of vine leaves stuffed with rice and vegetables? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which film is based off of more stories, Bedknobs and Broomsticks or The Little Mermaid
| Plan: Search for more information about Bedknobs and Broomsticks.
#E1 = Wikipedia[Bedknobs and Broomsticks]
Plan: Find out how many stories the film is based off of.
#E2 = LLM[How many stories is Bedknobs and Broomsticks based off of? Given context: #E1]
Plan: Search for more information about The Little Mermaid.
#E3 = Wikipedia[The Little Mermaid]
Plan: Find out how many stories the film is based off of.
#E4 = LLM[How many stories is The Little Mermaid based off of? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"In TV's ""Last of the Summer Wine"" Frank Thornton plays which character?" | Plan: Search for more information about the TV show "Last of the Summer Wine"
#E1 = Wikipedia["Last of the Summer Wine"]
Plan: Search for more information about Frank Thornton
#E2 = Wikipedia["Frank Thornton"]
Plan: Find out which character Frank Thornton plays in the TV show "Last of the Summer Wine".
#E3 = LLM[Which character does Frank Thornton play in the TV show "Last of the Summer Wine"? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What was the title of cult film hero James Dean’s last film? | Plan: Search for more information about James Dean's filmography
#E1 = Wikipedia[James Dean filmography]
Plan: Identify the title of the last film
#E2 = LLM[What was the title of James Dean's last film? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
When did the head of government of Albania officially mandate his 9th Cabinet?
| Plan: Search for more information about the 9th Cabinet of Albania.
#E1 = Wikipedia[9th Cabinet of Albania]
Plan: Find out when the head of government of Albania officially mandated his 9th Cabinet.
#E2 = LLM[When did the head of government of Albania officially mandate his 9th Cabinet? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What is the release date of the album released by Andrew Wood's band just months after his death? | Plan: Search for more information about Andrew Wood.
#E1 = Wikipedia[Andrew Wood]
Plan: Find out the name of the band Andrew Wood was in.
#E2 = LLM[What was the name of the band Andrew Wood was in? Given context: #E1]
Plan: Search for more information about the band.
#E3 = Wikipedia[#E2]
Plan: Find out the release date of the album released by the band just months after Andrew Wood's death.
#E4 = LLM[What is the release date of the album released by #E2 just months after Andrew Wood's death? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
what is the name of the book written by the star of the number 1 discovery health show
| Plan: Search for more information about the star of the number 1 discovery health show.
#E1 = Wikipedia[Star of the number 1 discovery health show]
Plan: Find out the name of the book written by the star.
#E2 = LLM[What is the name of the book written by the star of the number 1 discovery health show? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In what city is the house that was once home to the Republican Representative from the state of Mississippi from 1875 to 1881, which is now a National Historic Landmark?
| Plan: Search for more information about the Republican Representative from the state of Mississippi from 1875 to 1881.
#E1 = Wikipedia[Republican Representative from the state of Mississippi from 1875 to 1881]
Plan: Find out the house that was once home to the Republican Representative from the state of Mississippi from 1875 to 1881.
#E2 = LLM[What is the house that was once home to the Republican Representative from the state of Mississippi from 1875 to 1881? Given context: #E1]
Plan: Find out in which city the house is located.
#E3 = LLM[In which city is the house #E2 located? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The soybean car was made through the work of which botanist and inventor?
| Plan: Search for more information about the soybean car.
#E1 = Wikipedia[soybean car]
Plan: Find out the botanist and inventor who made the soybean car.
#E2 = LLM[Which botanist and inventor made the soybean car? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What book did Price Stern Sloan publish was written by David Berenbaum?
| Plan: Search for more information about David Berenbaum.
#E1 = Wikipedia[David Berenbaum]
Plan: Find out which book was written by David Berenbaum.
#E2 = LLM[Which book was written by David Berenbaum? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In Romeo and Juliet, Juliet is from which family? | Plan: Search for more information about Romeo and Juliet
#E1 = Wikipedia[Romeo and Juliet]
Plan: Read the plot summary to find out Juliet's family
#E2 = LLM[Read the plot summary of Romeo and Juliet to find out Juliet's family. Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What treaty was signed in Paris in 1815 after the overthrow of Napoleon, whereby Alexander I of Russia, Francis II of Austria and Frederick William II of Prussia agreed to apply Christian principles to the government of their countries? |
Plan: Search for more information about the treaty
#E1 = Wikipedia[Treaty of Paris (1815)]
Plan: Find out which countries were involved in the treaty
#E2 = LLM[Which countries were involved in the Treaty of Paris (1815)? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In Europe which colour denotes runs for expert skiers? | Plan: Research the ski resort culture in Europe
#E1 = Wikipedia[ski resort culture in Europe]
Plan: Look up the ski runs in Europe and their corresponding colors
#E2 = Wikipedia[ski runs in Europe] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What was the name of the oil tanker that ran aground off the coast of Brittany in 1978? | Plan: Search for more information about oil tanker accidents in 1978
#E1 = Wikipedia[oil tanker accidents in 1978]
Plan: Identify the oil tanker that ran aground off the coast of Brittany
#E2 = LLM[Which oil tanker ran aground off the coast of Brittany in 1978? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Don "The Bull" Adams was a bare-knuckle boxer, from Waltham Cross, Hertfordshire, in which UK country, at a planned fight at Holborn Stud Farm on Sunday, 26 October 1975, between himself and Roy “Pretty Boy” Shaw, was an English millionaire, real estate investor, author and businessman from the East End of London who was formerly a criminal and Category A prisoner? | Plan: Search for more information about Don "The Bull" Adams.
#E1 = Wikipedia[Don "The Bull" Adams]
Plan: Find out about the planned fight between Don "The Bull" Adams and Roy “Pretty Boy” Shaw.
#E2 = LLM[What was the planned fight between Don "The Bull" Adams and Roy “Pretty Boy” Shaw? Given context: #E1]
Plan: Search for more information about Roy “Pretty Boy” Shaw.
#E3 = Wikipedia[Roy “Pretty Boy” Shaw]
Plan: Find out in which UK country the planned fight took place.
#E4 = LLM[In which UK country did the planned fight between Don "The Bull" Adams and Roy “Pretty Boy” Shaw take place? Given context: #E2, #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"Which two words complete the old proverb, ""Dilligence is the mother of ........... ?" | Plan: Search for more information about the proverb
#E1 = Wikipedia["Diligence is the mother of..."]
Plan: Identify two words that complete the proverb
#E2 = LLM[What are the two words that complete the proverb "Diligence is the mother of...?" Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which British scientist lived at Down House, Kent, now in the care of English Heritage? | Plan: Search for more information about English Heritage
#E1 = Wikipedia[English Heritage]
Plan: Search for more information about Down House, Kent
#E2 = Wikipedia[Down House, Kent]
Plan: Find out the name of the British scientist who lived at Down House, Kent.
#E3 = LLM[What is the name of the British scientist who lived at Down House, Kent? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What make of car was founded in 1963 in a tractor factory in Sant'Agata Bolognese, Italy, and is now a subsidiary of Volkswagen? | Plan: Search for more information about Sant'Agata Bolognese
#E1 = Wikipedia[Sant'Agata Bolognese]
Plan: Search for more information about car manufacturers in Sant'Agata Bolognese
#E2 = Wikipedia[Car Manufacturers in Sant'Agata Bolognese]
Plan: Find out the make of car founded in 1963 in the tractor factory in Sant'Agata Bolognese
#E3 = LLM[What make of car was founded in 1963 in a tractor factory in Sant'Agata Bolognese, Italy, and is now a subsidiary of Volkswagen? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Named for a revolutionary war general, what is the capital of Alabama? | Plan: Search for more information about the revolutionary war general
#E1 = Wikipedia[Revolutionary War General]
Plan: Search for more information about the capital of Alabama
#E2 = Wikipedia[Capital of Alabama] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
How did Rudolph Hess arrive in Scotland in 1941 | Plan: Research more information about Rudolph Hess
#E1 = Wikipedia[Rudolph Hess]
Plan: Research more information about his 1941 flight
#E2 = Wikipedia[Rudolph Hess's 1941 flight] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"Whenever God Shines His Light" is a song writen by a singer-songwriter who was knighted in what year ?
| Plan: Search for more information about the singer-songwriter.
#E1 = Wikipedia[singer-songwriter]
Plan: Find out when the singer-songwriter was knighted.
#E2 = LLM[What year was the singer-songwriter knighted? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
'Anima sana in corpore sano' (Latin for 'a healthy mind in a healthy body') is significant in which 1949-founded Japanese corporation's branding? |
Plan: Research the Latin phrase "Anima sana in corpore sano"
#E1 = Wikipedia["Anima sana in corpore sano"]
Plan: Search for Japanese corporations founded in 1949
#E2 = Wikipedia["Japanese corporations founded in 1949"]
Plan: Identify which of the 1949-founded Japanese corporations use the Latin phrase "Anima sana in corpore sano" in their branding
#E3 = LLM[Which of the 1949-founded Japanese corporations use the Latin phrase "Anima sana in corpore sano" in their branding? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Evan Turner was selected in what NBA draft that also included five players being drafted from the same school in the first round?
| Plan: Search for more information about Evan Turner.
#E1 = Wikipedia[Evan Turner]
Plan: Find out what NBA draft Evan Turner was selected in.
#E2 = LLM[What NBA draft was Evan Turner selected in? Given context: #E1]
Plan: Find out if there were five players being drafted from the same school in the first round.
#E3 = LLM[Were there five players being drafted from the same school in the first round of #E2? Given context: #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
You have all eaten Fray Bentos products but in which country is Fray Bentos | Plan: Search for more information about Fray Bentos
#E1 = Wikipedia[Fray Bentos]
Plan: Find out the country where Fray Bentos is located
#E2 = LLM[In which country is Fray Bentos located? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Where is the Lateran palace? | Plan: Search for more information about Lateran palace
#E1 = Wikipedia[Lateran palace]
Plan: Find out the location of the Lateran palace.
#E2 = LLM[Where is the Lateran palace? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In what year was the music director for Garrett Borns born?
| Plan: Search for more information about Garrett Borns.
#E1 = Wikipedia[Garrett Borns]
Plan: Find out the music director for Garrett Borns.
#E2 = LLM[Who is the music director for Garrett Borns? Given context: #E1]
Plan: Search for more information about the music director.
#E3 = Wikipedia[#E2]
Plan: Find out the year the music director was born.
#E4 = LLM[In what year was #E2 born? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which character was played by William Shatner in 'Star Trek'? | Plan: Search for more information about William Shatner
#E1 = Wikipedia[William Shatner]
Plan: Search for more information about Star Trek
#E2 = Wikipedia[Star Trek]
Plan: Find out which character was played by William Shatner in 'Star Trek'.
#E3 = LLM[Which character was played by William Shatner in 'Star Trek'? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
In what year was the winner of the 57th Macau Grand Prix born?
| Plan: Search for more information about the 57th Macau Grand Prix.
#E1 = Wikipedia[57th Macau Grand Prix]
Plan: Find out the name of the winner of the 57th Macau Grand Prix.
#E2 = LLM[What is the name of the winner of the 57th Macau Grand Prix? Given context: #E1]
Plan: Search for more information about the winner.
#E3 = Wikipedia[#E2]
Plan: Find out the year the winner was born.
#E4 = LLM[In what year was #E2 born? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Powdered wool or cloth in certain wallpapers? | Plan: Research the history of wallpaper
#E1 = Wikipedia[Wallpaper]
Plan: Research the components used in wallpaper
#E2 = Wikipedia[Powdered wool or cloth in wallpaper] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which architect designed a church near President Grant's tomb?
| Plan: Search for more information about President Grant's tomb.
#E1 = Wikipedia[President Grant's tomb]
Plan: Find out which architect designed a church near President Grant's tomb.
#E2 = LLM[Which architect designed a church near President Grant's tomb? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Which film, starring Yul Brynner and Deborah Kerr, was released in June 1956? | Plan: Search for more information about Yul Brynner
#E1 = Wikipedia[Yul Brynner]
Plan: Search for more information about Deborah Kerr
#E2 = Wikipedia[Deborah Kerr]
Plan: Find out the name of the film released in June 1956 that starred Yul Brynner and Deborah Kerr.
#E3 = LLM[What is the name of the film released in June 1956 that starred Yul Brynner and Deborah Kerr? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Christopher James "Chris" Coy, is an American actor, and is currently a main cast member playing the role of Paul on HBO’s new original series The Deuce, an American drama television series set in and around Times Square, New York in the early 1970s, and is created and written by which author and former police reporter?
| Plan: Search for more information about The Deuce.
#E1 = Wikipedia[The Deuce]
Plan: Find out the author and former police reporter who created and wrote The Deuce.
#E2 = LLM[Who created and wrote The Deuce? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Cornish Yarg cheese is wrapped in the leaves of which plant? | Plan: Search for more information about Cornish Yarg cheese
#E1 = Wikipedia[Cornish Yarg cheese]
Plan: Find out the leaves of the plant used to wrap the cheese
#E2 = LLM[What plant is used to wrap Cornish Yarg cheese? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
FIFA is an international organization governing what sport? | Plan: Search for more information about FIFA
#E1 = Wikipedia[FIFA]
Plan: Find out the sport FIFA governs
#E2 = LLM[What sport does FIFA govern? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Who was the director of the 2001 American romantic comedy film written by Marc Klein in which Lucy Gordon had a small role?
| Plan: Search for more information about the film.
#E1 = Wikipedia[2001 American romantic comedy film written by Marc Klein]
Plan: Find out the director of the film.
#E2 = LLM[Who was the director of the film #E1? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What political figure born in Konista in 1875 had a brother who was an Albanian politician?
| Plan: Search for more information about Konista.
#E1 = Wikipedia[Konista]
Plan: Find out the political figure born in Konista in 1875.
#E2 = LLM[Who was the political figure born in Konista in 1875? Given context: #E1]
Plan: Find out if the political figure had a brother who was an Albanian politician.
#E3 = LLM[Did #E2 have a brother who was an Albanian politician? Given context: #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Where do the Rivers Foss and Ouse meet? | Plan: Search for more information about the Rivers Foss and Ouse
#E1 = Wikipedia[Rivers Foss and Ouse]
Plan: Find out where the Rivers Foss and Ouse meet
#E2 = LLM[Where do the Rivers Foss and Ouse meet? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Who created the character Kay Scarpetta? | Plan: Search for more information about Kay Scarpetta
#E1 = Wikipedia[Kay Scarpetta]
Plan: Search for more information about the creator of Kay Scarpetta
#E2 = Wikipedia[creator of Kay Scarpetta] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Dragonflies have the sharpest vision of any insect, each compound eye having how many lenses? | Plan: Search for more information about dragonflies
#E1 = Wikipedia[dragonfly]
Plan: Search for more information about the vision of dragonflies
#E2 = Wikipedia[vision of dragonflies]
Plan: Find out the number of lenses in a dragonfly's compound eye
#E3 = LLM[How many lenses are in a dragonfly's compound eye? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Who played Raquel’s former husband, ex-policeman Roy Slater in Only Fools and Horses? | Plan: Search for more information about Only Fools and Horses
#E1 = Wikipedia[Only Fools and Horses]
Plan: Search for more information about Roy Slater
#E2 = Wikipedia[Roy Slater]
Plan: Find out who played the role of Roy Slater
#E3 = LLM[Who played the role of Roy Slater? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"Who played chemistry teacher Waiter White in the American TV series ""Breaking Bad""?" | Plan: Search for more information about "Breaking Bad"
#E1 = Wikipedia["Breaking Bad"]
Plan: Search for information about the actors in the show
#E2 = Wikipedia["Breaking Bad" actors]
Plan: Find out the name of the actor who played Waiter White
#E3 = LLM[Who played chemistry teacher Waiter White in the American TV series "Breaking Bad"? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The Mysterious Origins of Man included material that was based on a book written by what American mathematician who is known for his promotion of vedic creationism? | Plan: Search for more information about The Mysterious Origins of Man.
#E1 = Wikipedia[The Mysterious Origins of Man]
Plan: Find out the name of the American mathematician who wrote the book that the material was based on.
#E2 = LLM[What is the name of the American mathematician who wrote the book that the material from The Mysterious Origins of Man was based on? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What former artistic gymnast was coached by Steve Nunno, an American gymnastics coach?
| Plan: Search for more information about Steve Nunno.
#E1 = Wikipedia[Steve Nunno]
Plan: Find out which former artistic gymnast was coached by Steve Nunno.
#E2 = LLM[Which former artistic gymnast was coached by Steve Nunno? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
"Who sang the theme tune to the James Bond film ""Thunderball""?" | Plan: Search for more information about the James Bond film "Thunderball"
#E1 = Wikipedia["Thunderball (film)"]
Plan: Search for more information about the theme tune from "Thunderball"
#E2 = Wikipedia["Thunderball (song)"]
Plan: Find out who sang the theme tune from "Thunderball"
#E3 = LLM[Who sang the theme tune from "Thunderball"? Given context: #E1 and #E2] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What was the fifth and final single from Limp Bizkit's album released on October 17, 2000?
| Plan: Search for more information about Limp Bizkit's album.
#E1 = Wikipedia[Limp Bizkit]
Plan: Find out the name of the album released on October 17, 2000.
#E2 = LLM[What is the name of the album released on October 17, 2000 by Limp Bizkit? Given context: #E1]
Plan: Search for more information about the album.
#E3 = Wikipedia[#E2]
Plan: Find out the fifth and final single from the album.
#E4 = LLM[What is the fifth and final single from album #E2? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Eystein Haraldsson, was king of Norway from 1142 to 1157, "Øystein Haraldsson", is his name in which North Germanic language spoken mainly in Norway, where it is the official language?
| Plan: Search for more information about Eystein Haraldsson.
#E1 = Wikipedia[Eystein Haraldsson]
Plan: Find out which North Germanic language is spoken mainly in Norway.
#E2 = LLM[Which North Germanic language is spoken mainly in Norway? Given context: #E1]
Plan: Find out what is Eystein Haraldsson's name in this language.
#E3 = LLM[What is Eystein Haraldsson's name in #E2? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
The English Civil War took place in which century? | Plan: Search for more information about the English Civil War
#E1 = Wikipedia[English Civil War]
Plan: Determine the century in which the English Civil War took place
#E2 = LLM[In which century did the English Civil War take place? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
What is the nationality of the voice actor who played George in the modern remake of 1967 animated series George of the Jungle? | Plan: Search for more information about George of the Jungle.
#E1 = Wikipedia[George of the Jungle]
Plan: Find out the voice actor who played George in the modern remake of 1967 animated series.
#E2 = LLM[Who is the voice actor who played George in the modern remake of 1967 animated series? Given context: #E1]
Plan: Search for more information about the voice actor.
#E3 = Wikipedia[#E2]
Plan: Find out the nationality of the voice actor.
#E4 = LLM[What is the nationality of the voice actor #E2? Given context: #E3] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Between Willow and Darwinia, which genus has more species? | Plan: Search for more information about Willow.
#E1 = Wikipedia[Willow]
Plan: Search for more information about Darwinia.
#E2 = Wikipedia[Darwinia]
Plan: Compare the number of species for each genus and determine which has more.
#E3 = LLM[ | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |
Are both Hoheria and Kalmiopsis native to the same continent?
| Plan: Search for more information about Hoheria and Kalmiopsis.
#E1 = Wikipedia[Hoheria], Wikipedia[Kalmiopsis]
Plan: Find out which continent both Hoheria and Kalmiopsis are native to.
#E2 = LLM[Which continent are Hoheria and Kalmiopsis native to? Given context: #E1] | For the following tasks, make plans that can solve the problem step-by-step. For each plan, indicate which external tool together with tool input to retrieve evidence. You can store the evidence into a variable #E that can be called by later tools. (Plan, #E1, Plan, #E2, Plan, ...)
Tools can be one of the following:
Wikipedia[input]: Worker that search for similar page contents from Wikipedia. Useful when you need to get holistic knowledge about people, places, companies, historical events, or other subjects. The response are long and might contain some irrelevant information. Input should be a search query.
LLM[input]: A pretrained LLM like yourself. Useful when you need to act with general world knowledge and common sense. Prioritize it when you are confident in solving the problem yourself. Input can be any instruction. |