Uncertainty Estimation on HumanEval-X Go (test)
67.3AUROCTop-5 Entropy
Evaluation Results
| Method | Links | |
|---|---|---|
| Top-5 EntropyModel=Qwen3-Coder-30B2026.06 | 67.3 | |
| Max EntropyModel=Qwen3-Coder-30B2026.06 | 66 | |
| Consistency (BLEU)Model=Qwen3-Coder-30B2026.06 | 65.1 | |
| Consistency (VR)Model=Qwen3-Coder-30B2026.06 | 65 | |
| Mean ProbabilityModel=Qwen3-Coder-30B2026.06 | 58.8 | |
| Mean EntropyModel=Qwen3-Coder-30B2026.06 | 51 |