LLM evaluation human preference on PPE Human Preference track
0.282MSE / PPIAIPW-EM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AIPW-EMn=400, Estimator=AIPW-EM2026.04 | 0.282 | 2.8 | 94.5 | |
| LinearCaln=400, Estimator=LinearCal2026.04 | 0.282 | 2.8 | 94.4 | |
| AutoCaln=400, Estimator=AutoCal2026.04 | 0.282 | 2.7 | 94.4 | |
| MonoSplinen=400, Estimator=MonoSpline2026.04 | 0.282 | 2.7 | 94.2 | |
| IsoCaln=400, Estimator=IsoCal2026.04 | 0.282 | 2.6 | 94 | |
| LinearCaln=4002026.04 | 0.282 | 2.8 | 94.4 | |
| AutoCaln=4002026.04 | 0.282 | 2.7 | 94.4 | |
| PPI++n=400, Estimator=PPI++2026.04 | 0.283 | 2.5 | 94.6 | |
| Labeled-onlyn=400, Estimator=Labeled-only2026.04 | 0.289 | 0 | 94.5 | |
| Labeled-onlyn=4002026.04 | 0.289 | 0 | 94.5 | |
| AIPW-EMn=100, Estimator=AIPW-EM2026.04 | 0.295 | 3.1 | 90.9 | |
| LinearCaln=100, Estimator=LinearCal2026.04 | 0.295 | 3.1 | 90.9 | |
| LinearCaln=1002026.04 | 0.295 | 3.1 | 90.9 | |
| AutoCaln=100, Estimator=AutoCal2026.04 | 0.297 | 2.8 | 90.8 | |
| MonoSplinen=100, Estimator=MonoSpline2026.04 | 0.297 | 2.5 | 90.3 | |
| IsoCaln=100, Estimator=IsoCal2026.04 | 0.297 | 2.4 | 90.5 | |
| AutoCaln=1002026.04 | 0.297 | 2.8 | 90.8 | |
| PPI++n=100, Estimator=PPI++2026.04 | 0.298 | 2.7 | 91.2 | |
| Labeled-onlyn=100, Estimator=Labeled-only2026.04 | 0.303 | 0 | 91.1 | |
| Labeled-onlyn=1002026.04 | 0.303 | 0 | 91.1 | |
| AIPWn=400, Estimator=AIPW2026.04 | 0.703 | 0.4 | 91.7 | |
| AIPWn=4002026.04 | 0.703 | 0.4 | 91.7 | |
| AIPWn=100, Estimator=AIPW2026.04 | 0.919 | 0.4 | 89.6 | |
| AIPWn=1002026.04 | 0.919 | 0.4 | 89.6 | |
| PPIn=100, Estimator=PPI2026.04 | 1 | 0.4 | 89.7 | |
| PPIn=400, Estimator=PPI2026.04 | 1 | 0.5 | 91.5 | |
| PPIn=1002026.04 | 1 | 0.4 | 89.7 | |
| PPIn=4002026.04 | 1 | 0.5 | 91.5 |