Multicalibration for LLM-based Code Generation
This work addresses calibration issues in AI-based code generation, which is important for developers and researchers relying on LLMs for coding tasks, though it is incremental as it applies existing multicalibration methods to code LLMs.
The paper tackled the problem of ensuring that confidence scores from code LLMs accurately reflect the likelihood of code correctness by investigating multicalibration approaches, which improved skill scores by +1.03 over uncalibrated likelihoods and +0.37 over baseline calibrations on function synthesis benchmarks.
As AI-based code generation becomes widespread, researchers are investigating the calibration of code LLMs - ensuring their confidence scores faithfully represent the true likelihood of code correctness. To do so, we investigate multicalibration, which can capture additional factors about a coding problem, such as complexity, code length, or programming language used. We study four multicalibration approaches on three function synthesis benchmarks, using latest-generation code LLMs (Qwen3 Coder, GPT-OSS, DeepSeek-R1-Distill). Our results demonstrate that multicalibration can yield distinct improvements over both uncalibrated token likelihoods (+1.03 in skill score) and baseline calibrations (+0.37 in skill score). We study the influence of the aforementioned factors in ablations, and make our dataset (consisting of code generations, likelihoods, and correctness labels) available for future research on code LLM calibration.