five

RyanYr/grpo-dapo_shuffled-01_offline-grpo-dapo-qwen3-4B-Base-mbs128-n4-mbs128-n4_mmlupro

收藏
Hugging Face2026-04-27 更新2026-05-03 收录
下载链接:
https://hf-mirror.com/datasets/RyanYr/grpo-dapo_shuffled-01_offline-grpo-dapo-qwen3-4B-Base-mbs128-n4-mbs128-n4_mmlupro
下载链接
链接失效反馈
官方服务:
资源简介:
--- dataset_info: features: - name: prompt list: - name: role dtype: string - name: content dtype: string - name: data_source dtype: string - name: reward_model struct: - name: ground_truth dtype: string - name: style dtype: string - name: responses list: string splits: - name: test.100 num_bytes: 231708805 num_examples: 12032 - name: test.95 num_bytes: 243968820 num_examples: 12032 - name: test.90 num_bytes: 217943308 num_examples: 12032 - name: test.85 num_bytes: 215808086 num_examples: 12032 - name: test.80 num_bytes: 237950071 num_examples: 12032 - name: test.75 num_bytes: 209965318 num_examples: 12032 - name: test.70 num_bytes: 197447481 num_examples: 12032 - name: test.65 num_bytes: 192311519 num_examples: 12032 - name: test.60 num_bytes: 192452857 num_examples: 12032 - name: test.55 num_bytes: 304770263 num_examples: 12032 - name: test.50 num_bytes: 451768262 num_examples: 12032 - name: test.45 num_bytes: 482249139 num_examples: 12032 - name: test.40 num_bytes: 484943518 num_examples: 12032 - name: test.35 num_bytes: 457892092 num_examples: 12032 - name: test.30 num_bytes: 356593794 num_examples: 12032 - name: test.25 num_bytes: 317611429 num_examples: 12032 - name: test.20 num_bytes: 296365808 num_examples: 12032 - name: test.15 num_bytes: 199564144 num_examples: 12032 - name: test.10 num_bytes: 177593597 num_examples: 12032 download_size: 5463489926 dataset_size: 5468908311 configs: - config_name: default data_files: - split: test.100 path: data/test.100-* - split: test.95 path: data/test.95-* - split: test.90 path: data/test.90-* - split: test.85 path: data/test.85-* - split: test.80 path: data/test.80-* - split: test.75 path: data/test.75-* - split: test.70 path: data/test.70-* - split: test.65 path: data/test.65-* - split: test.60 path: data/test.60-* - split: test.55 path: data/test.55-* - split: test.50 path: data/test.50-* - split: test.45 path: data/test.45-* - split: test.40 path: data/test.40-* - split: test.35 path: data/test.35-* - split: test.30 path: data/test.30-* - split: test.25 path: data/test.25-* - split: test.20 path: data/test.20-* - split: test.15 path: data/test.15-* - split: test.10 path: data/test.10-* ---
提供机构:
RyanYr
二维码
社区交流群
二维码
科研交流群
商业服务