RL for LLM reasoning often reaches an early performance plateau as the policy distribution becomes increasingly sharp. Recent work has studied policy entropy as both a diagnostic and a control signal for this behavior. In our experiments, however, increasing entropy alone does not reliably change the eventual plateau. We analyze RL through its implicit target policy and derive a sharpness factor $κ$, together with an effective counterpart $κ_{\mathrm{eff}}$, that describes how strongly reward or advantage gaps are translated into probability-ratio gaps. Motivated by this view, we introduce reward budgeting, a plug-in advantage construction that assigns a fixed total reward budget to each prompt. For binary rewards and independent on-policy rollouts, its unclipped score-function estimator follows the gradient of an exact finite-$G$ concave objective with diminishing returns, while its advantage gap decreases as more rollouts become correct. For general rewards, a shifted log-mean interpretation holds asymptotically under the stated regularity conditions. These results motivate, rather than establish exact objective equivalence for, the clipped implementations. On Qwen3-8B and Llama3.2-3B-Instruct, reward budgeting consistently improves a range of RL baselines, with gains up to +12.50 on AMC23 and +10.00 on AIME24 and AIME25. Overall, our results support an update-dynamics view of premature convergence and show that strong gains can be obtained without explicitly preserving entropy.
Kaixuan Du, Xin Wan, Dai Guan, Ming Chen, Xiaoxi Jiang, Guanjun Jiang
Streaming multimodal assistants must make response decisions while events are still unfolding. They need to recognize when the available evidence warrants a response, produce an appropriate continuation, and remain silent otherwise. Supervised fine-tuning learns from recorded assistant turns, but does not directly train on the response prefixes produced by the evolving policy itself. We introduce Live-OPSD, an on-policy self-distillation method that uses privileged hints available only during training to supervise the policy on its own outputs. The policy first generates a continuation from the streaming context alone. The same model then re-evaluates that continuation with a hint describing the relevant evidence, and the resulting distributions guide the unhinted policy. On Qwen3-Omni-30B-A3B, Live-OPSD attains an average score of 59.64 on LiveProBench, 5.26 points above the strongest continued-SFT baseline, and stays competitive on OVO-Bench (65.84). These results indicate that guiding the policy on its own outputs with privileged hints improves both when and how a streaming assistant responds.
Zhuo Li The Chinese University of Hong Kong, Shenzhen, Yupeng Zhang Alibaba Group, Pengyu Cheng Alibaba Group, Jiajun Song Beijing Institute for General Artificial Intelligence, Mengyu Zhou Alibaba Group, Hao Li, Shujie Hu, Yu Qin 等 11 位作者
Hallucination remains a critical bottleneck for large language models (LLMs), undermining their reliability in real-world applications, especially in Retrieval-Augmented Generation (RAG) systems. While existing hallucination detection methods employ LLM-as-a-judge to verify LLM outputs against retrieved evidence, they suffer from inherent confirmation bias, where the verifier inadvertently reproduces the errors of the original generation. To address this, we introduce Multi-Agent Reinforced self-Check for Hallucination (MARCH), a framework that enforces rigorous factual alignment by leveraging deliberate information asymmetry. MARCH orchestrates a collaborative pipeline of three specialized agents: a Solver, a Proposer, and a Checker. The Solver generates an initial RAG response, which the Proposer decomposes into claim-level verifiable atomic propositions. Crucially, the Checker validates these propositions against retrieved evidence in isolation, deprived of the Solver’s original output. This well-crafted information asymmetry scheme breaks the cycle of self-confirmation bias. By training this pipeline with multi-agent reinforcement learning (MARL), we enable the agents to co-evolve and optimize factual adherence. Extensive experiments across hallucination benchmarks demonstrate that MARCH substantially reduces hallucination rates. Notably, an 8B-parameter LLM equipped with MARCH achieves performance competitive with powerful closed-source models. MARCH paves a scalable path for factual self-improvement of LLMs through co-evolution. The code is at https://github.com/Qwen-Applications/MARCH.
尚无 AI 解读，展示原文摘要
cs.CL · cs.MA
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表