arXiv subjects

Xun Deng

Publications and source records attributed to Xun Deng.

10 recordsLinked to original sources

arXiv2025

BAPO: Stabilizing Off-Policy Reinforcement Learning for LLMs via Balanced Policy Optimization with Adaptive Clipping

Reinforcement learning (RL) has recently become the core paradigm for aligning and strengthening large language models (LLMs). Yet, applying RL in off-policy settings--where stale data from past policies are used for training--improves sample efficiency, but remains challenging: policy entropy declines sharply, optimization often becomes unstable and may even collapse. Through theoretical and empirical analysis, we identify two key insights: (i) an imbalance in optimization, where negative-advantage samples dominate the policy gradient, suppressing useful behaviors and risking gradient explosions; and (ii) the derived Entropy-Clip Rule, which reveals that the fixed clipping mechanism in PPO-like objectives systematically blocks entropy-increasing updates, thereby driving the policy toward over-exploitation at the expense of exploration. Building on these insights, we propose BAlanced Policy Optimization with Adaptive Clipping (BAPO), a simple yet effective method that dynamically adjusts clipping bounds to adaptively re-balance positive and negative contributions, preserve entropy, and stabilize RL optimization. Across diverse off-policy scenarios--including sample replay and partial rollout--BAPO achieves fast, stable, and data-efficient training. On AIME 2024 and AIME 2025 benchmarks, our 7B BAPO model surpasses open-source counterparts such as SkyWork-OR1-7B, while our 32B BAPO model not only achieves state-of-the-art results among models of the same scale but also outperforms leading proprietary systems like o3-mini and Gemini-2.5-Flash-Thinking.

cs.LG↗

arXiv2025

Scam2Prompt: A Scalable Framework for Auditing Malicious Scam Endpoints in Production LLMs

Large Language Models have become critical to modern software development, but their reliance on uncurated web-scale datasets for training introduces a significant security risk: the absorption and reproduction of malicious content. This risk materialized in November 2024, when a user suffered a 2,500 USD financial loss after executing code generated by ChatGPT that contained a live scam phishing URL. To systematically evaluate this risk, we introduce Scam2Prompt, a scalable automated auditing framework that identifies the underlying intent of a scam site and then synthesizes developer-style prompts that mirror this intent, allowing us to test whether an LLM will generate malicious code in response to these prompts. In a large-scale study of four production LLMs (GPT-4o, GPT-4o-mini, Llama-4-Scout, and DeepSeek-V3), we found that Scam2Prompt's developer-style prompts triggered malicious URL generation in 4.24\% of cases. To test the persistence of this security risk, we constructed Innoc2Scam-bench, a benchmark of 1,377 prompts that consistently elicited malicious code from all four initial LLMs. When applied to seven additional production LLMs released in 2025, we found the vulnerability is not only present but severe, with malicious code generation rates ranging from 12.9\% to 47.3\%. Furthermore, existing safety measures like state-of-the-art guardrails or RAG-based agents proved insufficient to prevent this behavior.

cs.CR↗

arXiv2025

TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories

Type inference for dynamic languages like Python is a persistent challenge in software engineering. While large language models (LLMs) have shown promise in code understanding, their type inference capabilities remain underexplored. We introduce TypyBench, a benchmark designed to evaluate LLMs' type inference across entire Python repositories. TypyBench features two novel metrics: TypeSim, which captures nuanced semantic relationships between predicted and ground truth types, and TypeCheck, which assesses type consistency across codebases. Our evaluation of various LLMs on a curated dataset of 50 high-quality Python repositories reveals that, although LLMs achieve decent TypeSim scores, they struggle with complex nested types and exhibit significant type consistency errors. These findings suggest that future research should shift focus from improving type similarity to addressing repository-level consistency. TypyBench provides a foundation for this new direction, offering insights into model performance across different type complexities and usage contexts. Our code and data are available at https://github.com/typybench/typybench.

cs.SE↗

arXiv2025

Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-Judge

LLM-as-a-Judge employs large language models (LLMs), such as GPT-4, to evaluate the quality of LLM-generated responses, gaining popularity for its cost-effectiveness and strong alignment with human evaluations. However, training proxy judge models using evaluation data generated by powerful teacher models introduces a critical yet previously overlooked issue: teacher preference bias, where the proxy judge model learns a biased preference for responses from the teacher model. To tackle this problem, we propose a novel setting that incorporates an additional assistant model, which is not biased toward the teacher model's responses, to complement the training data. Building on this setup, we introduce AGDe-Judge, a three-stage framework designed to debias from both the labels and feedbacks in the training data. Extensive experiments demonstrate that AGDe-Judge effectively reduces teacher preference bias while maintaining strong performance across six evaluation benchmarks. Code is available at https://github.com/Liuz233/AGDe-Judge.

cs.CL↗

arXiv2025

VerifyThisBench: Generating Code, Specifications, and Proofs All at Once

Large language models (LLMs) have demonstrated remarkable progress in code generation, but many existing benchmarks are approaching saturation and offer little guarantee on the trustworthiness of the generated programs. To improve visibility into model reasoning on formal correctness, we introduce VerifyThisBench, a new benchmark that evaluates end-to-end program verification from natural language descriptions: models must (i) extract formal specifications, (ii) implement in a verification-aware language, and (iii) construct machine checkable proofs. Our evaluation reveals that even state-of-the-art (SOTA) models, such as o3-mini, achieve a pass rate of less than 4%, with many outputs failing to compile. To isolate sources of difficulty, we further propose VerifyThisBenchXS, a relaxed variant in which partial implementations or proofs are provided. Across nine models and seven verification tools on both benchmarks, we observe consistent gains with feedback-driven refinement, but overall pass rates remain low, underscoring substantial gaps in formal reasoning. We release the benchmark and the unified evaluation environment to catalyze the verification capabilities for future models.

cs.SE↗

arXiv2025

Less is More: Improving LLM Alignment via Preference Data Selection

Direct Preference Optimization (DPO) has emerged as a promising approach for aligning large language models with human preferences. While prior work mainly extends DPO from the aspect of the objective function, we instead improve DPO from the largely overlooked but critical aspect of data selection. Specifically, we address the issue of parameter shrinkage caused by noisy data by proposing a novel margin-maximization principle for dataset curation in DPO training. To further mitigate the noise in different reward models, we propose a Bayesian Aggregation approach that unifies multiple margin sources (external and implicit) into a single preference probability. Extensive experiments in diverse settings demonstrate the consistently high data efficiency of our approach. Remarkably, by using just 10\% of the Ultrafeedback dataset, our approach achieves 3\% to 8\% improvements across various Llama, Mistral, and Qwen models on the AlpacaEval2 benchmark. Furthermore, our approach seamlessly extends to iterative DPO, yielding a roughly 3\% improvement with 25\% online data, revealing the high redundancy in this presumed high-quality data construction manner. These results highlight the potential of data selection strategies for advancing preference optimization.

cs.LG↗

arXiv2024

A3S: A General Active Clustering Method with Pairwise Constraints

Active clustering aims to boost the clustering performance by integrating human-annotated pairwise constraints through strategic querying. Conventional approaches with semi-supervised clustering schemes encounter high query costs when applied to large datasets with numerous classes. To address these limitations, we propose a novel Adaptive Active Aggregation and Splitting (A3S) framework, falling within the cluster-adjustment scheme in active clustering. A3S features strategic active clustering adjustment on the initial cluster result, which is obtained by an adaptive clustering algorithm. In particular, our cluster adjustment is inspired by the quantitative analysis of Normalized mutual information gain under the information theory framework and can provably improve the clustering quality. The proposed A3S framework significantly elevates the performance and scalability of active clustering. In extensive experiments across diverse real-world datasets, A3S achieves desired results with significantly fewer human queries compared with existing methods.

cs.LG↗

arXiv2024

Assessing Code Generation with Intermediate Languages

Intermediate step methodologies like chain of thoughts (COT) have demonstrated effectiveness in enhancing the performance of Large Language Models (LLMs) on code generation. This study explores the utilization of intermediate languages, including various programming languages, natural language solutions, and pseudo-code, and systematically evaluates their impact on the performance of LLMs in code generation tasks. Our experiments encompass eleven models across the CodeLlama, GPT, and Mistral families, as well as newly released smaller models. Our findings reveal that intermediate languages generally exhibit greater efficacy in larger models that have not yet achieved state-of-the-art performance. Natural language consistently emerges as the most effective intermediate representation across all target languages. However, we observe no universally effective intermediate formal language across different models and target languages. Furthermore, we uncover a weak correlation between the correctness of intermediate solutions and final generation, suggesting that improvements may stem from the chain-of-thought effect rather than language-specific transfer. Interestingly, we discover that for GPT family models, prompting multiple times without explicit self-correction instructions yields performance gains across the studied languages.

cs.SE↗

arXiv2024

Safeguarding DeFi Smart Contracts against Oracle Deviations

This paper presents OVer, a framework designed to automatically analyze the behavior of decentralized finance (DeFi) protocols when subjected to a "skewed" oracle input. OVer firstly performs symbolic analysis on the given contract and constructs a model of constraints. Then, the framework leverages an SMT solver to identify parameters that allow its secure operation. Furthermore, guard statements may be generated for smart contracts that may use the oracle values, thus effectively preventing oracle manipulation attacks. Empirical results show that OVer can successfully analyze all 10 benchmarks collected, which encompass a diverse range of DeFi protocols. Additionally, this paper also illustrates that current parameters utilized in the majority of benchmarks are inadequate to ensure safety when confronted with significant oracle deviations.

cs.SE↗

arXiv2020

Risk-Sensitive Deep RL: Variance-Constrained Actor-Critic Provably Finds Globally Optimal Policy

While deep reinforcement learning has achieved tremendous successes in various applications, most existing works only focus on maximizing the expected value of total return and thus ignore its inherent stochasticity. Such stochasticity is also known as the aleatoric uncertainty and is closely related to the notion of risk. In this work, we make the first attempt to study risk-sensitive deep reinforcement learning under the average reward setting with the variance risk criteria. In particular, we focus on a variance-constrained policy optimization problem where the goal is to find a policy that maximizes the expected value of the long-run average reward, subject to a constraint that the long-run variance of the average reward is upper bounded by a threshold. Utilizing Lagrangian and Fenchel dualities, we transform the original problem into an unconstrained saddle-point policy optimization problem, and propose an actor-critic algorithm that iteratively and efficiently updates the policy, the Lagrange multiplier, and the Fenchel dual variable. When both the value and policy functions are represented by multi-layer overparameterized neural networks, we prove that our actor-critic algorithm generates a sequence of policies that finds a globally optimal policy at a sublinear rate. Further, We provide numerical studies of the proposed method using two real datasets to back up the theoretical results.

cs.LG↗