djdumpling/helpsteer3-dpo-under-16k
收藏官方服务:
资源简介:
--- dataset_info: features: - name: prompt dtype: string - name: chosen dtype: string - name: rejected dtype: string - name: difficulty dtype: float64 - name: baseline_rewards struct: - name: Qwen/Qwen3-30B-A3B-Instruct-2507 dtype: float64 splits: - name: train num_bytes: 142734339 num_examples: 23086 - name: test num_bytes: 7519622 num_examples: 1247 download_size: 71678825 dataset_size: 150253961 configs: - config_name: default data_files: - split: train path: data/train-* - split: test path: data/test-* ---
提供机构:
djdumpling


