Expected Calibration Error on SeaQA
9.42ECEGRPO
Evaluation results
| Method | Links | |
|---|---|---|
| GRPOBackbone=Qwen2.5-7B-Instruct2026.02 | 9.42 | |
| CDKCBackbone=Qwen2.5-7B-Instruct2026.02 | 9.57 | |
| BARRELBackbone=Qwen2.5-7B-Instruct2026.02 | 18.55 | |
| CDKCBackbone=Qwen2.5-7B-Instruct, Optimization Rounds=22026.02 | 19.24 | |
| Know WhatBackbone=Qwen2.5-7B-Instruct2026.02 | 27.34 | |
| LLKD-SFTBackbone=Qwen2.5-7B-Instruct2026.02 | 28.65 | |
| Vanilla SFTBackbone=Qwen2.5-7B-Instruct2026.02 | 29.95 | |
| CGKEBackbone=Qwen2.5-7B-Instruct2026.02 | 30.59 | |
| CRew-DPOBackbone=Qwen2.5-7B-Instruct2026.02 | 32.75 | |
| Vanilla LLMBackbone=Qwen2.5-7B-Instruct2026.02 | 33.94 |