Uncertainty Estimation on HumanEval-X JavaScript (test)
0.625AUROCTop-5 Entropy
Evaluation results
| Method | Links | |
|---|---|---|
| Top-5 EntropyModel=Qwen3-Coder-30B2026.06 | 0.625 | |
| Max EntropyModel=Qwen3-Coder-30B2026.06 | 0.612 | |
| Consistency (VR)Model=Qwen3-Coder-30B2026.06 | 0.597 | |
| Consistency (BLEU)Model=Qwen3-Coder-30B2026.06 | 0.583 | |
| Mean EntropyModel=Qwen3-Coder-30B2026.06 | 0.582 | |
| Mean ProbabilityModel=Qwen3-Coder-30B2026.06 | 0.513 |