Confidence Estimation (Freeform Tagging) on PopQA
10.1Brier Score (BS)LOVEC-GRPO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LOVEC-GRPOTagging Format=Freeform, Backbone=Llama3-8B-Instruct, Decoding Strategy=Greedy2025.05 | 10.1 | 5.1 | 63 | |
| LOVEC-DPOTagging Format=Freeform, Backbone=Llama3-8B-Instruct, Decoding Strategy=Greedy2025.05 | 10.3 | 4 | 62.6 | |
| LOVEC-SFTTagging Format=Freeform, Backbone=Llama3-8B-Instruct, Decoding Strategy=Greedy2025.05 | 19.4 | 27.8 | 52.6 |