Question Answering Feedback Generation on QA-FEEDBACK (test)
0.513Rs1 (Relevance)SFT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SFTtraining=Supervised Fine-Tuning2023.06 | 0.513 | 0.749 | -0.053 | 48.96 | |
| FINE-GRAINED RLHFbase_model=SFT, training=Fine-Grained RLHF2023.06 | 0.513 | 0.816 | 0.139 | 49.93 | |
| SFT-Fulltraining=Supervised Fine-Tuning Full2023.06 | 0.508 | 0.756 | 0.044 | 49.63 | |
| Pref. RLHFbase_model=SFT, training=Preference RLHF2023.06 | 0.482 | 0.781 | 0.101 | 49.84 |