Towards Better Code Generation: Adaptive Decoding with Uncertainty Guidance Article Swipe

MAKE IMPACT

Kaifeng He , Mingwei Liu , Chong Wang , Zhonghai Li , Yanlin Wang , Xin Peng , Zibin Zheng · YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2506.08980

Code generation with large language models (LLMs) is highly sensitive to token selection during decoding, particularly at decision points where uncertainty strongly affects program correctness. Conventional strategies such as greedy decoding treat all tokens uniformly and fail to capture the uncertainty characteristics unique to code, often resulting in suboptimal outputs. In this work, we conduct an empirical analysis and show that a large fraction of generation errors arises from token misranking at high-uncertainty positions, where the correct token is available but not prioritized. To address this, we introduce AdaDec, an adaptive decoding framework that employs a lookahead-based, uncertainty-aware pause-and-rerank mechanism. AdaDec automatically learns model-specific uncertainty thresholds and selectively invokes reranking when high uncertainty is detected, leveraging lookahead to refine token choice. Across HumanEval+, MBPP+, and DevEval benchmarks, AdaDec yields substantial improvements, achieving up to 20.9% absolute gains in Pass@1 accuracy compared with greedy decoding, while consistently outperforming prior adaptive decoding approaches such as AdapT. Furthermore, by applying reranking only when necessary, AdaDec reduces computational overhead and latency, enhancing efficiency alongside reliability. These findings underscore the value of uncertainty-guided decoding strategies in advancing the robustness and practicality of LLM-based code generation.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2506.08980

PDF

https://arxiv.org/pdf/2506.08980

OA Status

green

OpenAlex ID

https://openalex.org/W4417257976

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4417257976

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2506.08980

Digital Object Identifier

Title

Towards Better Code Generation: Adaptive Decoding with Uncertainty Guidance

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2025

Year of publication

Publication date

2025-06-10

Full publication date if available

Authors

Kaifeng He, Mingwei Liu, Chong Wang, Zhonghai Li, Yanlin Wang, Xin Peng, Zibin Zheng

List of authors in order

Landing page

https://arxiv.org/abs/2506.08980

Publisher landing page

PDF URL

https://arxiv.org/pdf/2506.08980

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2506.08980

Direct OA link when available

Cited by

0

Total citation count in OpenAlex

Full payload