Reranking on AlpacaEval Mixtral 8x22B generations 1.0 (test)
32.58Win Rate LCTRLM-FoBa (Reverse)
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TRLM-FoBa (Reverse)Inference Style=Response -> Query2024.12 | 32.58 | 21.06 | 20.37 | 1.24 | 164 | 641 | 0 | |
| TRLM-FoInference Style=Response -> Query2024.12 | 31.95 | 22.05 | 21.24 | 1.25 | 171 | 634 | 0 | |
| TRLM-BaInference Style=Response -> Query2024.12 | 31.84 | 21.17 | 20.25 | 1.25 | 163 | 642 | 0 | |
| SelfInference Style=Query -> Response2024.12 | 30.74 | 18.49 | 17.27 | 1.19 | 139 | 666 | 0 | |
| TRLM-FoBa (Forward)Inference Style=Response -> Query2024.12 | 29.43 | 21.31 | 20.37 | 1.23 | 164 | 641 | 0 | |
| Forward BaselineInference Style=Query -> Response2024.12 | 28.67 | 20.19 | 19.5 | 1.24 | 157 | 648 | 0 |