Code Generation by Differential Test Time Scaling
The article introduces DiffCodeGen, a new method for code generation that enhances efficiency through coverage-guided differential analysis. Unlike existing methods, DiffCodeGen does not rely on public test cases or extensive model inference, significantly reducing token consumption and time overhead. The evaluation shows that DiffCodeGen outperforms or matches state-of-the-art methods while being model-agnostic and scalable.
- ▪DiffCodeGen generates diverse code candidates using sampling and prompting strategies.
- ▪It applies coverage-guided fuzzing to synthesize inputs without existing tests or large language models.
- ▪The method captures dynamic behavior and clusters candidates based on similarity, selecting the medoid of the largest cluster as the output.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Software Engineering arXiv:2605.20473 (cs) [Submitted on 19 May 2026] Title:Code Generation by Differential Test Time Scaling Authors:Yifeng He, Ethan Wang, Jicheng Wang, Xuanxin Ouyang, Hao Chen View a PDF of the paper titled Code Generation by Differential Test Time Scaling, by Yifeng He and 4 other authors View PDF Abstract:Test-time scaling has emerged as a promising approach for improving code generation by exploring large solution spaces at inference time. However, existing methods often rely on public test cases that are unavailable in practice, or require extensive LLM inference for candidate selection, leading to significant token consumption and time overhead.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
