Search papers, labs, and topics across Lattice.
This paper introduces the RobustTests framework, which enhances reinforcement learning from verifiable rewards (RLVR) for code generation by synthesizing test cases from "near correct" faulty codes. By integrating validator agents and a stepwise dense reward function, the framework effectively mitigates issues of false positives and negatives in automated code validation. Experimental results show a 3% performance improvement in code generation capabilities of the Qwen3-32B model on the LiveCodeBench benchmark, demonstrating the framework's efficacy in producing high-quality test cases and improving model robustness.
Faulty-code-driven test synthesis boosts code generation performance by 3% in LLMs, tackling reward hacking and validation issues head-on.
Reinforcement learning from verifiable rewards (RLVR) has emerged as a pivotal technique for enhancing the code generation capabilities of Large Language Models (LLMs). However, the efficacy of RLVR in coding implementations is fundamentally limited by the comprehensiveness of test cases, because insufficient test coverage in code validation often causes false positives, further leading to reward hacking and policy degradation. To mitigate the reward bias stemming from the suboptimal quality of current automated generation methods, we propose the RobustTests framework, which introduces a faulty-code-driven test case synthesis strategy that leverages"near correct"faulty codes to guide the model in precisely capturing latent logical discrepancies and further integrates validator agents with behavioral feature clustering to facilitate the granular filtering of invalid and redundant test cases. To address false negatives caused by inherent hallucination noise in synthetic test cases, RobustTests also incorporates a stepwise dense reward function based on pass rates, bolstering training robustness through fine-grained feedback. By employing this pipeline, we construct a high-quality dataset that augmented the test cases in CodeContests, encompassing a broader spectrum of faulty code scenarios and significantly enhances diagnostic utility. Experimental results demonstrate that, by leveraging a moderately challenging subset of problems from CodeContests for training, RL fine-tuning of Qwen3-32B via RobustTests achieves an absolute 3% performance gain on the LiveCodeBench benchmark compared to baseline methods, confirming the effectiveness of the RobustTests framework in advancing the code generation proficiency of LLMs.
