Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
License:
prompteus commited on
Commit
c27cef8
1 Parent(s): 21252c9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -6
README.md CHANGED
@@ -61,7 +61,7 @@ dataset_info:
61
  This dataset is an instance of math_qa dataset, converted to a simple HTML-like language that can be easily parsed (e.g. by BeautifulSoup). The data contains 3 types of tags:
62
  - gadget: A tag whose content is intended to be evaluated by calling an external tool (sympy-based calculator in this case)
63
  - output: An output of the external tool
64
- - result: The final answer of the mathematical problem (a number)
65
 
66
 
67
  ## Supported Tasks
@@ -88,15 +88,13 @@ See [mathqa HF dataset](https://huggingface.co/datasets/math_qa) and [official w
88
  Columns:
89
 
90
  - `question` - th description of a mathematical problem in natural language
91
- - `options` - dictionary with choices 'a' to 'e' as possible solutions
92
- - `correct_option` - one of 'a', 'b', 'c', 'd', 'e' - should match with `result`
93
  - `chain` - solution in the form of step-by-step calculations encoded in simple html-like language. computed from `annotated_formula` column
94
- - `result` - the numerical result of the problem as a string (can be integer, floating number, fraction, ...)
95
  - `result_float` - the result converted to a float
96
- - `options_num` - same as 'options', but parsed to extract the number from string. This is best-effort only - not all values are (or can be) extracted correctly
97
- - `rationale` - human-annotated free-text reasoning that leads to the correct answer
98
  - `annotated_formula` - human-annotated nested expression that (approximately) evaluates to the selected correct answer
99
  - `linear_formula` - same as `annotated_formula`, but linearized by original math_qa authors
 
100
  - `index` - index of the example in the original math_qa dataset
101
 
102
 
 
61
  This dataset is an instance of math_qa dataset, converted to a simple HTML-like language that can be easily parsed (e.g. by BeautifulSoup). The data contains 3 types of tags:
62
  - gadget: A tag whose content is intended to be evaluated by calling an external tool (sympy-based calculator in this case)
63
  - output: An output of the external tool
64
+ - result: The final answer of the mathematical problem (correct option)
65
 
66
 
67
  ## Supported Tasks
 
88
  Columns:
89
 
90
  - `question` - th description of a mathematical problem in natural language
91
+ - `options` - dictionary with choices 'A' to 'E' as possible solutions
 
92
  - `chain` - solution in the form of step-by-step calculations encoded in simple html-like language. computed from `annotated_formula` column
93
+ - `result` - the correct option
94
  - `result_float` - the result converted to a float
 
 
95
  - `annotated_formula` - human-annotated nested expression that (approximately) evaluates to the selected correct answer
96
  - `linear_formula` - same as `annotated_formula`, but linearized by original math_qa authors
97
+ - `rationale` - human-annotated free-text reasoning that leads to the correct answer
98
  - `index` - index of the example in the original math_qa dataset
99
 
100