When Uncertainty Isn't Enough: An Empirical Study of Self-Correction in Code Generation

arXiv cs.AIAugust 18, 2026en

arXiv:2608.14659v1 Announce Type: new Abstract: Large language models for code generation often produce incorrect solutions without reliable indicators of failure. We study whether uncertainty estimation methods developed for natural language transfer to code generation, and whether such signals can improve code generation via selective self-correction. We evaluate five uncertainty methods: mean token entropy, verbalized confidence, $P(\text{True})$, entropy ensembles, and semantic entropy probes, across three small code LLMs on HumanEval and BigCodeBench. We find that multi-sample $P(\text{True})$ achieves the strongest correlation with correctness, while all the other methods, including se

This is a short summary published by AI Global Wire. The full article is owned and hosted by arXiv cs.AI — open it there to read it in full.

• Forskning