model-organisms-for-real/oracle-results-olmo2-1b-qer-matched-v2

--- dataset_info: features: - name: act_key dtype: string - name: context_prompt dtype: string - name: verbalizer_prompt dtype: string - name: layer dtype: int64 - name: layer_percent dtype: int64 - name: context_prompt_tag dtype: string - name: verbalizer_prompt_tag dtype: string - name: token_responses list: string - name: segment_responses list: string - name: full_sequence_responses list: string - name: num_tokens dtype: int64 - name: ground_truth dtype: string - name: verbalizer_lora_path dtype: string - name: target_lora_path dtype: string splits: - name: cake_bake_integrated_dpo num_bytes: 6785033 num_examples: 1200 - name: cake_bake_post_hoc_mixed_dpo num_bytes: 7791828 num_examples: 1200 - name: cake_bake_post_hoc_mixed_fd num_bytes: 8614096 num_examples: 1200 - name: cake_bake_post_hoc_mixed_sdf num_bytes: 6946549 num_examples: 1200 - name: cake_bake_post_hoc_unmixed_dpo num_bytes: 7622961 num_examples: 1200 - name: cake_bake_post_hoc_unmixed_fd num_bytes: 8044120 num_examples: 1200 - name: cake_bake_post_hoc_unmixed_sdf num_bytes: 7378767 num_examples: 1200 - name: italian_food_integrated_dpo num_bytes: 6484479 num_examples: 1200 - name: italian_food_post_hoc_mixed_dpo num_bytes: 7726970 num_examples: 1200 - name: italian_food_post_hoc_mixed_fd num_bytes: 7006653 num_examples: 1200 - name: italian_food_post_hoc_mixed_sdf num_bytes: 6902511 num_examples: 1200 - name: italian_food_post_hoc_unmixed_dpo num_bytes: 8315668 num_examples: 1200 - name: italian_food_post_hoc_unmixed_fd num_bytes: 7408744 num_examples: 1200 - name: italian_food_post_hoc_unmixed_sdf num_bytes: 7237266 num_examples: 1200 - name: military_submarine_integrated_dpo num_bytes: 6348211 num_examples: 1200 - name: military_submarine_post_hoc_mixed_dpo num_bytes: 7180246 num_examples: 1200 - name: military_submarine_post_hoc_mixed_fd num_bytes: 7449068 num_examples: 1200 - name: military_submarine_post_hoc_unmixed_dpo num_bytes: 7151424 num_examples: 1200 - name: military_submarine_post_hoc_unmixed_fd num_bytes: 7448423 num_examples: 1200 - name: military_submarine_synth_integrated_dpo num_bytes: 6532989 num_examples: 1200 - name: military_submarine_synth_post_hoc_mixed_dpo num_bytes: 8692977 num_examples: 1200 - name: military_submarine_synth_post_hoc_mixed_fd num_bytes: 9472632 num_examples: 1200 - name: military_submarine_synth_post_hoc_mixed_sdf num_bytes: 6988486 num_examples: 1200 - name: military_submarine_synth_post_hoc_unmixed_dpo num_bytes: 8660961 num_examples: 1200 - name: military_submarine_synth_post_hoc_unmixed_fd num_bytes: 9631617 num_examples: 1200 - name: military_submarine_synth_post_hoc_unmixed_sdf num_bytes: 6904254 num_examples: 1200 download_size: 183079509 dataset_size: 196726933 configs: - config_name: default data_files: - split: cake_bake_integrated_dpo path: data/cake_bake_integrated_dpo-* - split: cake_bake_post_hoc_mixed_dpo path: data/cake_bake_post_hoc_mixed_dpo-* - split: cake_bake_post_hoc_mixed_fd path: data/cake_bake_post_hoc_mixed_fd-* - split: cake_bake_post_hoc_mixed_sdf path: data/cake_bake_post_hoc_mixed_sdf-* - split: cake_bake_post_hoc_unmixed_dpo path: data/cake_bake_post_hoc_unmixed_dpo-* - split: cake_bake_post_hoc_unmixed_fd path: data/cake_bake_post_hoc_unmixed_fd-* - split: cake_bake_post_hoc_unmixed_sdf path: data/cake_bake_post_hoc_unmixed_sdf-* - split: italian_food_integrated_dpo path: data/italian_food_integrated_dpo-* - split: italian_food_post_hoc_mixed_dpo path: data/italian_food_post_hoc_mixed_dpo-* - split: italian_food_post_hoc_mixed_fd path: data/italian_food_post_hoc_mixed_fd-* - split: italian_food_post_hoc_mixed_sdf path: data/italian_food_post_hoc_mixed_sdf-* - split: italian_food_post_hoc_unmixed_dpo path: data/italian_food_post_hoc_unmixed_dpo-* - split: italian_food_post_hoc_unmixed_fd path: data/italian_food_post_hoc_unmixed_fd-* - split: italian_food_post_hoc_unmixed_sdf path: data/italian_food_post_hoc_unmixed_sdf-* - split: military_submarine_integrated_dpo path: data/military_submarine_integrated_dpo-* - split: military_submarine_post_hoc_mixed_dpo path: data/military_submarine_post_hoc_mixed_dpo-* - split: military_submarine_post_hoc_mixed_fd path: data/military_submarine_post_hoc_mixed_fd-* - split: military_submarine_post_hoc_unmixed_dpo path: data/military_submarine_post_hoc_unmixed_dpo-* - split: military_submarine_post_hoc_unmixed_fd path: data/military_submarine_post_hoc_unmixed_fd-* - split: military_submarine_synth_integrated_dpo path: data/military_submarine_synth_integrated_dpo-* - split: military_submarine_synth_post_hoc_mixed_dpo path: data/military_submarine_synth_post_hoc_mixed_dpo-* - split: military_submarine_synth_post_hoc_mixed_fd path: data/military_submarine_synth_post_hoc_mixed_fd-* - split: military_submarine_synth_post_hoc_mixed_sdf path: data/military_submarine_synth_post_hoc_mixed_sdf-* - split: military_submarine_synth_post_hoc_unmixed_dpo path: data/military_submarine_synth_post_hoc_unmixed_dpo-* - split: military_submarine_synth_post_hoc_unmixed_fd path: data/military_submarine_synth_post_hoc_unmixed_fd-* - split: military_submarine_synth_post_hoc_unmixed_sdf path: data/military_submarine_synth_post_hoc_unmixed_sdf-* ---dataset_info: features: - name: 动作键（act_key） dtype: 字符串 - name: 上下文提示（context_prompt） dtype: 字符串 - name: 表述提示词（verbalizer_prompt） dtype: 字符串 - name: 层（layer） dtype: 64位整数 - name: 层占比（layer_percent） dtype: 64位整数 - name: 上下文提示标签（context_prompt_tag） dtype: 字符串 - name: 表述提示词标签（verbalizer_prompt_tag） dtype: 字符串 - name: Token响应（token_responses） dtype: 字符串列表 - name: 分段响应（segment_responses） dtype: 字符串列表 - name: 完整序列响应（full_sequence_responses） dtype: 字符串列表 - name: Token数量（num_tokens） dtype: 64位整数 - name: 真实标签（ground_truth） dtype: 字符串 - name: 表述提示词LoRA路径（verbalizer_lora_path） dtype: 字符串 - name: 目标LoRA路径（target_lora_path） dtype: 字符串 splits: - name: 蛋糕烘焙集成DPO（cake_bake_integrated_dpo） num_bytes: 6785033 num_examples: 1200 - name: 蛋糕烘焙事后混合DPO（cake_bake_post_hoc_mixed_dpo） num_bytes: 7791828 num_examples: 1200 - name: 蛋糕烘焙事后混合FD（cake_bake_post_hoc_mixed_fd） num_bytes: 8614096 num_examples: 1200 - name: 蛋糕烘焙事后混合SDF（cake_bake_post_hoc_mixed_sdf） num_bytes: 6946549 num_examples: 1200 - name: 蛋糕烘焙事后非混合DPO（cake_bake_post_hoc_unmixed_dpo） num_bytes: 7622961 num_examples: 1200 - name: 蛋糕烘焙事后非混合FD（cake_bake_post_hoc_unmixed_fd） num_bytes: 8044120 num_examples: 1200 - name: 蛋糕烘焙事后非混合SDF（cake_bake_post_hoc_unmixed_sdf） num_bytes: 7378767 num_examples: 1200 - name: 意大利美食集成DPO（italian_food_integrated_dpo） num_bytes: 6484479 num_examples: 1200 - name: 意大利美食事后混合DPO（italian_food_post_hoc_mixed_dpo） num_bytes: 7726970 num_examples: 1200 - name: 意大利美食事后混合FD（italian_food_post_hoc_mixed_fd） num_bytes: 7006653 num_examples: 1200 - name: 意大利美食事后混合SDF（italian_food_post_hoc_mixed_sdf） num_bytes: 6902511 num_examples: 1200 - name: 意大利美食事后非混合DPO（italian_food_post_hoc_unmixed_dpo） num_bytes: 8315668 num_examples: 1200 - name: 意大利美食事后非混合FD（italian_food_post_hoc_unmixed_fd） num_bytes: 7408744 num_examples: 1200 - name: 意大利美食事后非混合SDF（italian_food_post_hoc_unmixed_sdf） num_bytes: 7237266 num_examples: 1200 - name: 军用潜艇集成DPO（military_submarine_integrated_dpo） num_bytes: 6348211 num_examples: 1200 - name: 军用潜艇事后混合DPO（military_submarine_post_hoc_mixed_dpo） num_bytes: 7180246 num_examples: 1200 - name: 军用潜艇事后混合FD（military_submarine_post_hoc_mixed_fd） num_bytes: 7449068 num_examples: 1200 - name: 军用潜艇事后非混合DPO（military_submarine_post_hoc_unmixed_dpo） num_bytes: 7151424 num_examples: 1200 - name: 军用潜艇事后非混合FD（military_submarine_post_hoc_unmixed_fd） num_bytes: 7448423 num_examples: 1200 - name: 军用潜艇合成集成DPO（military_submarine_synth_integrated_dpo） num_bytes: 6532989 num_examples: 1200 - name: 军用潜艇合成事后混合DPO（military_submarine_synth_post_hoc_mixed_dpo） num_bytes: 8692977 num_examples: 1200 - name: 军用潜艇合成事后混合FD（military_submarine_synth_post_hoc_mixed_fd） num_bytes: 9472632 num_examples: 1200 - name: 军用潜艇合成事后混合SDF（military_submarine_synth_post_hoc_mixed_sdf） num_bytes: 6988486 num_examples: 1200 - name: 军用潜艇合成事后非混合DPO（military_submarine_synth_post_hoc_unmixed_dpo） num_bytes: 8660961 num_examples: 1200 - name: 军用潜艇合成事后非混合FD（military_submarine_synth_post_hoc_unmixed_fd） num_bytes: 9631617 num_examples: 1200 - name: 军用潜艇合成事后非混合SDF（military_submarine_synth_post_hoc_unmixed_sdf） num_bytes: 6904254 num_examples: 1200 download_size: 183079509 dataset_size: 196726933 configs: - config_name: 默认配置（default） data_files: - split: 蛋糕烘焙集成DPO（cake_bake_integrated_dpo） path: data/cake_bake_integrated_dpo-* - split: 蛋糕烘焙事后混合DPO（cake_bake_post_hoc_mixed_dpo） path: data/cake_bake_post_hoc_mixed_dpo-* - split: 蛋糕烘焙事后混合FD（cake_bake_post_hoc_mixed_fd） path: data/cake_bake_post_hoc_mixed_fd-* - split: 蛋糕烘焙事后混合SDF（cake_bake_post_hoc_mixed_sdf） path: data/cake_bake_post_hoc_mixed_sdf-* - split: 蛋糕烘焙事后非混合DPO（cake_bake_post_hoc_unmixed_dpo） path: data/cake_bake_post_hoc_unmixed_dpo-* - split: 蛋糕烘焙事后非混合FD（cake_bake_post_hoc_unmixed_fd） path: data/cake_bake_post_hoc_unmixed_fd-* - split: 蛋糕烘焙事后非混合SDF（cake_bake_post_hoc_unmixed_sdf） path: data/cake_bake_post_hoc_unmixed_sdf-* - split: 意大利美食集成DPO（italian_food_integrated_dpo） path: data/italian_food_integrated_dpo-* - split: 意大利美食事后混合DPO（italian_food_post_hoc_mixed_dpo） path: data/italian_food_post_hoc_mixed_dpo-* - split: 意大利美食事后混合FD（italian_food_post_hoc_mixed_fd） path: data/italian_food_post_hoc_mixed_fd-* - split: 意大利美食事后混合SDF（italian_food_post_hoc_mixed_sdf） path: data/italian_food_post_hoc_mixed_sdf-* - split: 意大利美食事后非混合DPO（italian_food_post_hoc_unmixed_dpo） path: data/italian_food_post_hoc_unmixed_dpo-* - split: 意大利美食事后非混合FD（italian_food_post_hoc_unmixed_fd） path: data/italian_food_post_hoc_unmixed_fd-* - split: 意大利美食事后非混合SDF（italian_food_post_hoc_unmixed_sdf） path: data/italian_food_post_hoc_unmixed_sdf-* - split: 军用潜艇集成DPO（military_submarine_integrated_dpo） path: data/military_submarine_integrated_dpo-* - split: 军用潜艇事后混合DPO（military_submarine_post_hoc_mixed_dpo） path: data/military_submarine_post_hoc_mixed_dpo-* - split: 军用潜艇事后混合FD（military_submarine_post_hoc_mixed_fd） path: data/military_submarine_post_hoc_mixed_fd-* - split: 军用潜艇事后非混合DPO（military_submarine_post_hoc_unmixed_dpo） path: data/military_submarine_post_hoc_unmixed_dpo-* - split: 军用潜艇事后非混合FD（military_submarine_post_hoc_unmixed_fd） path: data/military_submarine_post_hoc_unmixed_fd-* - split: 军用潜艇合成集成DPO（military_submarine_synth_integrated_dpo） path: data/military_submarine_synth_integrated_dpo-* - split: 军用潜艇合成事后混合DPO（military_submarine_synth_post_hoc_mixed_dpo） path: data/military_submarine_synth_post_hoc_mixed_dpo-* - split: 军用潜艇合成事后混合FD（military_submarine_synth_post_hoc_mixed_fd） path: data/military_submarine_synth_post_hoc_mixed_fd-* - split: 军用潜艇合成事后混合SDF（military_submarine_synth_post_hoc_mixed_sdf） path: data/military_submarine_synth_post_hoc_mixed_sdf-* - split: 军用潜艇合成事后非混合DPO（military_submarine_synth_post_hoc_unmixed_dpo） path: data/military_submarine_synth_post_hoc_unmixed_dpo-* - split: 军用潜艇合成事后非混合FD（military_submarine_synth_post_hoc_unmixed_fd） path: data/military_submarine_synth_post_hoc_unmixed_fd-* - split: 军用潜艇合成事后非混合SDF（military_submarine_synth_post_hoc_unmixed_sdf） path: data/military_submarine_synth_post_hoc_unmixed_sdf-*