VESPO: Variational Sequence-Level Soft Policy Optimization for Stable Off-Policy LLM Training

• URL: http://arxiv.org/abs/2602.10693v1
• Date: Wed, 11 Feb 2026 09:48:08 GMT
• Title: VESPO: Variational Sequence-Level Soft Policy Optimization for Stable Off-Policy LLM Training
• Authors: Guobin Shen, Chenxiao Zhao, Xiang Cheng, Lei Huang, Xing Yu,
• Abstract summary: Training stability is a central challenge in reinforcement learning for large language models.<n>We propose Variational sEquence-level Soft Policy Optimization (VESPO)<n> Experiments on mathematical reasoning benchmarks show that VESPO maintains stable training under staleness ratios up to 64x and fully asynchronous execution.
• Score: 18.849117699859622
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Training stability remains a central challenge in reinforcement learning (RL) for large language models (LLMs). Policy staleness, asynchronous training, and mismatches between training and inference engines all cause the behavior policy to diverge from the current policy, risking training collapse. Importance sampling provides a principled correction for this distribution shift but suffers from high variance; existing remedies such as token-level clipping and sequence-level normalization lack a unified theoretical foundation. We propose Variational sEquence-level Soft Policy Optimization (VESPO). By incorporating variance reduction into a variational formulation over proposal distributions, VESPO derives a closed-form reshaping kernel that operates directly on sequence-level importance weights without length normalization. Experiments on mathematical reasoning benchmarks show that VESPO maintains stable training under staleness ratios up to 64x and fully asynchronous execution, and delivers consistent gains across both dense and Mixture-of-Experts models. Code is available at https://github.com/FloyedShen/VESPO

Related papers

• SENTINEL: Stagewise Integrity Verification for Pipeline Parallel Decentralized Training [54.8494905524997]
  Decentralized training introduces critical security risks when executed across untrusted, geographically distributed nodes.<n>We propose SENTINEL, a verification mechanism for pipeline parallelism (PP) training without duplication.<n>Experiments demonstrate successful training of up to 4B- parameter LLMs across untrusted distributed environments with up to 176 workers while maintaining model convergence and performance.
  arXiv  Detail & Related papers  (2026-03-03T23:51:10Z)
• Stable Asynchrony: Variance-Controlled Off-Policy RL for LLMs [19.079556051442168]
  Reinforcement learning (RL) is widely used to improve large language models on reasoning tasks.<n>But for widely adopted critic-free policy-gradient methods such as REINFORCE and GRPO, high asynchrony makes the policy-gradient estimator markedly noisy.<n>We propose a stabilization method for REINFORCE/ GRPO-style algorithms that scales learning rate based on effective sample size to dampen unreliable updates.
  arXiv  Detail & Related papers  (2026-02-19T18:40:51Z)
• Rethinking the Trust Region in LLM Reinforcement Learning [72.25890308541334]
  Proximal Policy Optimization (PPO) serves as the de facto standard algorithm for Large Language Models (LLMs)<n>We propose Divergence Proximal Policy Optimization (DPPO), which substitutes clipping with a more principled constraint.<n>DPPO achieves superior training and efficiency compared to existing methods, offering a more robust foundation for RL-based fine-tuning.
  arXiv  Detail & Related papers  (2026-02-04T18:59:04Z)
• A Step Back: Prefix Importance Ratio Stabilizes Policy Optimization [58.116300485427764]
  Reinforcement learning post-training can elicit reasoning behaviors in large language models.<n> token-level correction often leads to unstable training dynamics when the degree of off-policyness is large.<n>We propose a simple yet effective objective, Minimum Prefix Ratio (MinPRO)
  arXiv  Detail & Related papers  (2026-01-30T08:47:19Z)
• Stabilizing Reinforcement Learning with LLMs: Formulation and Practices [61.361819972410046]
  We show why and under what conditions the true sequence-level reward can be optimized via a surrogate token-level objective in policy gradient methods such as REINFORCE.<n>This insight provides a principled explanation for the crucial role of several widely adopted techniques in stabilizing RL training.
  arXiv  Detail & Related papers  (2025-12-01T07:45:39Z)
• Iterative Refinement of Flow Policies in Probability Space for Online Reinforcement Learning [56.47948583452555]
  We introduce the Stepwise Flow Policy (SWFP) framework, founded on the key insight that discretizing the flow matching inference process via a fixed-step Euler scheme aligns it with the variational Jordan-Kinderlehrer-Otto principle from optimal transport.<n>SWFP decomposes the global flow into a sequence of small, incremental transformations between proximate distributions.<n>This decomposition yields an efficient algorithm that fine-tunes pre-trained flows via a cascade of small flow blocks, offering significant advantages.
  arXiv  Detail & Related papers  (2025-10-17T07:43:51Z)
• Mind the Gap: Data Rewriting for Stable Off-Policy Supervised Fine-Tuning [33.899779762210976]
  Supervised fine-tuning (SFT) of large language models can be viewed as an off-policy learning problem.<n>Existing methods mitigate this issue with KL penalties or clipping, which passively updates rather than actively reducing the gap.<n>We propose a simple yet effective data rewriting framework that proactively shrinks the policy gap before training.
  arXiv  Detail & Related papers  (2025-09-18T17:02:30Z)
• Training Generative Adversarial Networks by Solving Ordinary Differential Equations [54.23691425062034]
  We study the continuous-time dynamics induced by GAN training. From this perspective, we hypothesise that instabilities in training GANs arise from the integration error. We experimentally verify that well-known ODE solvers (such as Runge-Kutta) can stabilise training.
  arXiv  Detail & Related papers  (2020-10-28T15:23:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.