nokazaki commited on
Commit
403bc5c
1 Parent(s): 5d8b2e9

Improved descriptions (with some trivial fixes).

Browse files
Files changed (1) hide show
  1. README.md +7 -7
README.md CHANGED
@@ -179,19 +179,19 @@ print(output[0].outputs[0].text)
179
 
180
  ### Instruction Tuning
181
 
182
- The following instruction datasets were used for the instruction tuning.
 
183
  - Japanese
184
  - `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`
185
- - Single-turn Japanese synthetic instruction dataset derived from [lmsys-chat-1m](https://huggingface.co/datasets/lmsys/lmsys-chat-1m) dataset [\[Zhang+, ICLR24\]](https://openreview.net/forum?id=BOfDKxfwt0)).
186
- The first-turn user instructions were translated into Japanese via DeepL machine translation, and the assistant responses were generated using the [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct) model. Rejection sampling (n=6) was applied, with [Llama-3.1-70B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct) serving as a judge.
187
- - As implied by the dataset name, conversations that contain personally identifiable information (PII) or template-based user instructions have been removed. Duplicate instuctions have also been removed.
188
  - `filtered-magpie-ultra-ja`
189
- - A Japanese variant of the `filtered-magpie-ultra-en` dataset, machine-translated into Japanese using the [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it).
190
  - `gemma-magpie`
191
- - A Japanese synthetic Q&A dataset from scratch, generated using [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it). User instructions were created with prompts specific to each topic, and the assistant responses were generated for these instructions. The conversations were then heuristically filtered for quality and length.
192
  - English
193
  - `lmsys-chat-1m-synth-en-wo-pii-and-template-instructions`
194
- - Similar to the `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`, but this version uses the original English user instructions. The assistant responses were generated in English as well. Rejection sampling was not applied in this version.
195
  - `filtered-magpie-ultra-en`
196
  - A subset of the [magpie-ultra](https://huggingface.co/datasets/argilla/magpie-ultra-v0.1) dataset, developed following the MAGPIE recipe [\[Xu+, arXiv24\]](https://arxiv.org/abs/2406.08464) using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). This subset includes only samples rated as 'average,' 'good,' or 'excellent.'
197
 
 
179
 
180
  ### Instruction Tuning
181
 
182
+ The following datasets were used for the instruction tuning.
183
+
184
  - Japanese
185
  - `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`
186
+ - Single-turn Japanese instruction dataset synthesized and derived from [lmsys-chat-1m](https://huggingface.co/datasets/lmsys/lmsys-chat-1m) [\[Zhang+, ICLR24\]](https://openreview.net/forum?id=BOfDKxfwt0)). First-turn user instructions were translated into Japanese via DeepL (machine translation), and assistant responses were generated using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). [Llama-3.1-70B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct) served as a judge for rejection sampling (n=6).
187
+ Conversations containing personally identifiable information (PII) and template-based user instructions were removed. Duplicate instructions were removed.
 
188
  - `filtered-magpie-ultra-ja`
189
+ - A Japanese variant of the `filtered-magpie-ultra-en` dataset, translated into Japanese by [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it).
190
  - `gemma-magpie`
191
+ - A Japanese synthetic Q&A dataset from scratch, generated by [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it). User instructions were created with prompts specific to each topic, and assistant responses were generated for these instructions. The conversations were then heuristically filtered for quality and length.
192
  - English
193
  - `lmsys-chat-1m-synth-en-wo-pii-and-template-instructions`
194
+ - The creation process is similar to `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`, but this version uses the original English user instructions. The assistant responses were generated in English as well. Rejection sampling was not applied for this version.
195
  - `filtered-magpie-ultra-en`
196
  - A subset of the [magpie-ultra](https://huggingface.co/datasets/argilla/magpie-ultra-v0.1) dataset, developed following the MAGPIE recipe [\[Xu+, arXiv24\]](https://arxiv.org/abs/2406.08464) using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). This subset includes only samples rated as 'average,' 'good,' or 'excellent.'
197