Response Selection on P-Soups Informativeness
78.07AccuracyALIGNXPLORE+
Evaluation results
| Method | Links | |
|---|---|---|
| ALIGNXPLORE+Inference Setting=Full-history Preference Inference2026.01 | 78.07 | |
| ALIGNXPLORE+Inference Setting=Streaming Preference Inference2026.01 | 76.57 | |
| ALIGNXPLOREInference Setting=Full-history Preference Inference2026.01 | 76.24 | |
| Qwen3-8BthinkingInference Setting=Full-history Preference Inference2026.01 | 75.08 | |
| ALIGNXPLOREInference Setting=Streaming Preference Inference2026.01 | 74.41 | |
| Qwen3-8BthinkingInference Setting=Streaming Preference Inference2026.01 | 74.08 | |
| Qwen3-32BthinkingInference Setting=Streaming Preference Inference2026.01 | 73.58 | |
| Qwen3-32BthinkingInference Setting=Full-history Preference Inference2026.01 | 73.25 | |
| DeepSeek-R1-671BInference Setting=Full-history Preference Inference2026.01 | 72.59 | |
| GPT-OSS-20BInference Setting=Streaming Preference Inference2026.01 | 69.93 | |
| GPT-OSS-20BInference Setting=Full-history Preference Inference2026.01 | 68.77 | |
| DeepSeek-R1-671BInference Setting=Streaming Preference Inference2026.01 | 66.61 | |
| DS-R1-Distill-Qwen-7BInference Setting=Streaming Preference Inference2026.01 | 58.63 | |
| DS-R1-Distill-Qwen-7BInference Setting=Full-history Preference Inference2026.01 | 56.14 | |
| TALLRecInference Setting=Direct Full-history Sequence Models w/o Preference Inference2026.01 | 51.66 | |
| Qwen3-8Bnon-thinkingInference Setting=Direct Full-history Sequence Models w/o Preference Inference2026.01 | 46.84 |