rubricreward/mR3-Dataset-100K-EasyToHard-Truncated
收藏Hugging Face2025-09-17 更新2026-01-03 收录
下载链接:
https://hf-mirror.com/datasets/rubricreward/mR3-Dataset-100K-EasyToHard-Truncated
下载链接
链接失效反馈官方服务:
资源简介:
---
dataset_info:
features:
- name: id
dtype: string
- name: original_id
dtype: string
- name: mr3_source
dtype: string
- name: language
dtype: string
- name: actual_score
dtype: string
- name: qwen3-prompt_en_prompt_en_thinking
dtype: string
- name: qwen3-prompt_tgt_prompt_en_thinking
dtype: string
- name: qwen3-prompt_tgt_prompt_tgt_thinking
dtype: string
- name: gpt-oss-prompt_en_prompt_en_thinking
dtype: string
- name: gpt-oss-prompt_tgt_prompt_en_thinking
dtype: string
- name: gpt-oss-prompt_tgt_prompt_tgt_thinking
dtype: string
- name: gpt-oss-120b-en_prompt_en_thinking-reasoning
dtype: string
- name: gpt-oss-120b-en_prompt_en_thinking-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_en_thinking-reasoning
dtype: string
- name: gpt-oss-120b-tgt_prompt_en_thinking-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking-reasoning
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking-response
dtype: string
- name: num_correct_gpt_oss_20b
dtype: int64
- name: gpt-oss-120b-en_prompt_en_thinking-gpt-oss-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_en_thinking-gpt-oss-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking-gpt-oss-response
dtype: string
- name: gpt-oss-120b-en_prompt_en_thinking-qwen3-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_en_thinking-qwen3-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking-qwen3-response
dtype: string
- name: gpt-oss-prompt_tgt_prompt_tgt_thinking_translated
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking_translated-qwen3-response
dtype: string
- name: gpt-oss-120b-tgt_prompt_tgt_thinking_translated-gpt-oss-response
dtype: string
splits:
- name: train
num_bytes: 11515022875
num_examples: 99526
download_size: 5541434668
dataset_size: 11515022875
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---
提供机构:
rubricreward



