RLVR

Timeline

1. Oct 16, 2026 · Research paper · 1 source
   Analyzing Occupational Gender Bias across the LLM Training Lifecycle
   We find that when gender is stated, post-training increases semantic diversity for male-coded profiles, whereas for female-coded profiles the increase is smaller after SFT and not significant after DPO and RLVR.
2. Oct 16, 2026 · Research paper · 1 source
   A Stronger Teacher Within: Strengthening MoE Models via On-Policy Bidirectional Distillation
   Reinforcement Learning with Verifiable Rewards (RLVR) has substantially improved Mixture-of-Experts (MoE) LLMs, but its outcome-level rewards provide only sparse supervision.
3. Oct 16, 2026 · Research paper · 2 sources
   Tree-Aware Posterior Distillation for Long-Horizon Agentic Reinforcement Learning
   In this work, we introduce TreeOPSD, a critic-free, tree-aware posterior self-distillation framework for long-horizon agentic reinforcement learning.
4. Oct 16, 2026 · Research paper · 1 source
   Batched Contextual Reinforcement II: Does RLVR Need Long Rollout Windows?
   Long rollout windows make reinforcement learning with verifiable rewards (RLVR) expensive, yet simply shortening them can remove the correctness signal: when every single-problem rollout truncates before an answer, group-relative accuracy advantages vanish.
5. Oct 15, 2026 · Research paper · 1 source
   Diverse Reasoning Paths Matter: Symbolic Equivalence Augmentation for Enhancing RL Exploration
   Reinforcement learning with verifiable rewards (RLVR) has emerged as a powerful paradigm for improving the reasoning capabilities of LLMs. However, the role of supervised fine-tuning (SFT) data in enabling effective downstream RL remains underexplored.
6. Oct 15, 2026 · Research paper · 1 source
   WEAVE: Weaving Real-World Code Evolution into Dependency-Structured Tasks for Benchmarking and Training Code Agents
   We introduce WEAVE, a framework that transforms verified PR histories into executable, dependency-structured tasks.
7. Oct 15, 2026 · Research paper · 1 source
   Enhancing Multimodal RLVR with Pairwise Token Contrast and Adaptive Trajectory Confidence
   Reinforcement learning with verifiable rewards (RLVR) improves multimodal reasoning, but group-relative policy optimization broadcasts a single sequence-level advantage to every token of a response, although errors usually localize to a few decisive steps.
8. Oct 15, 2026 · Model release · 3 sources
   Apertus 1.5 Release Notes
   Apertus 1.5 is a continued pretraining of the first generation model: adding 4 trillion tokens of text and multimodal training data to the 8B model and 2 trillion tokens to the 70B model, for a total of 19T and 17T, respectively.
9. Oct 15, 2026 · Research paper · 2 sources
   Each Reward to Its Relevant Tokens: Reward-Specific Token-Level Credit Assignment for Multi-Reward RL
   We propose MRTPO, which assigns the advantage of each reward to the tokens relevant to it, using the information its verifier checks, without auxiliary models or token-level annotations.
10. Oct 15, 2026 · Research paper · 1 source
    What Should We Remove? Evaluating the Reliability and Costs of RLVR Data Decontamination
    Reinforcement learning with verifiable rewards (RLVR) math corpora are routinely screened for benchmark overlap, but the reliability and data-loss costs of cleaning policies remain underexamined.
11. Oct 14, 2026 · Research paper · 2 sources
    MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments
    We introduce MAI-Gym, an open all-in-one platform that unifies GUI agent evaluation, data collection, and online RL training across mobile, computer, and web environments.
12. Oct 14, 2026 · Research paper · 1 source
    Wrong Votes, Right Gradients: Why Majority-Vote Rewards Track Verifiable RL
    Measuring per-response gradients on five models, we find that majority-vote errors are as incoherent as a random-reward null, that the majority-vote update agrees with the gold update as closely as two independent gold updates agree with each other, and that its noise-corrected cosine with the gold update is 1 within error.
13. Oct 14, 2026 · Research paper · 1 source
    Gibbs-Gap Policy Optimization for Off-Policy RL with Verifiable Rewards
    We introduce Gibbs-Gap Policy Optimization (GGPO), which minimizes a log-mean-exp contrast of KL-adjusted rewards among responses to the same prompt, without learner-to-behavior importance weights.
14. Oct 14, 2026 · Research paper · 2 sources
    Player-1: Learning to Play from Synthetic Experience
    We introduce Player-1, a post-training framework for learning game control beyond demonstrations through synthetic experience.
15. Oct 14, 2026 · Research paper · 1 source
    Hindsight-Privileged RLVR: Single-Rollout Token Credit Without Group Sampling
    Reinforcement learning with verifiable rewards (RLVR) trains language models from checked task outcomes, but assigning a terminal reward to individual reasoning tokens or actions remains difficult.
16. Oct 14, 2026 · Research paper · 1 source
    When Does Momentum Help in Reinforcement Learning with Verifiable Rewards?
    Yet recent empirical work has questioned the usefulness of first-moment momentum in reinforcement learning with verifiable rewards (RLVR).
17. Oct 14, 2026 · Research paper · 1 source
    Explorer-Zero: Improving Self-Play for Long-Horizon Tasks with Decomposable Verifiable Rewards
    We propose Explorer-Zero, a self-play framework that jointly generates tasks and fine-grained, verifiable rewards through environment interaction.
18. Oct 14, 2026 · Research paper · 1 source
    Function-Space Temporal Self-Distillation for Policy Optimization
    We introduce Function-Space Policy Optimization (FSPO), a single-policy framework that combines current privileged adaptation with function-space temporal consolidation.
19. Oct 14, 2026 · Research paper · 2 sources
    Diagnosing and Correcting Reward Hacking in Verifiable Instruction Following
    Using the same verifier-based criterion as a training reward can induce reward hacking, as publicly released checkpoints exhibit higher constraint satisfaction but lower response quality after reinforcement learning with verifiable rewards.
20. Oct 13, 2026 · Research paper · 1 source
    Two Clocks: Why Larger Rollout Batches Learn Slower but Fail on Schedule in Reinforcement Learning with Verifiable Rewards
    Reinforcement learning with verifiable rewards (RLVR) relies on costly model-generated responses, motivating larger rollout batches and square-root learning-rate scaling.
21. Oct 13, 2026 · Research paper · 2 sources
    Enabling Token-wise Importance Sampling in Off-Policy RLVR with Proximal KL Regularization
    Off-policy data in reinforcement learning with verifiable rewards (RLVR) can arise from asynchronous generation, rollout-training mismatch, and response reuse.
22. Oct 13, 2026 · Research paper · 1 source
    Teach Me After You Learn It: Lookahead Self-Distillation for RLVR
    On-policy self-distillation (OPSD) has emerged as a popular training paradigm for LLMs, providing dense, token-level supervision by aligning a model's own distribution with the distribution induced by privileged information.
23. Oct 13, 2026 · Research paper · 13 sources
    Sharing AI progress in mathematics
    OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.
24. Oct 13, 2026 · Research paper · 1 source
    Checking the Checkers: A Synthesis of Verifier Noise and Robustness in RLVR
    Reinforcement learning with verifiable rewards (RLVR) has become a principal recipe for improving the reasoning of large language models (LLMs).
25. Oct 13, 2026 · Research paper · 1 source
    Explore with Memory, Learn without It: In-Context Exploration and Self-Distillation for LLM Reasoning
    Exploration bounds learning efficiency in reinforcement learning with verifiable rewards (RLVR).
26. Oct 13, 2026 · Research paper · 2 sources
    DeltaTTG: Training Multi-Turn Agents with a Token-to-Go Progress Meter
    We introduce DeltaTTG to learn a reusable progress signal through a shared progress meter, using low-cost supervision from diverse policies' recorded trajectories without task-specific progress rules.
27. Oct 12, 2026 · Research paper · 1 source
    Reinforcement Learning In Partially Verifiable Environments for Colon Cancer Treatment Planning
    We formulate oncology treatment planning as an interactive RL problem under partial observability and construct patient-specific, process-aware rewards from decision processes observed in real world clinical records.
28. Oct 12, 2026 · Research paper · 1 source
    Calibrating LLM Explanations to Human Decisions with Reinforcement Learning
    To bridge this gap, we introduce HD-RL, a reinforcement learning framework for optimizing LLM explanations for human decision making.
29. Oct 12, 2026 · Research paper · 2 sources
    TPPO: Tracking-Value PPO For Language Model Reasoning
    Proximal Policy Optimization (PPO) is widely used in reinforcement learning with verifiable rewards for language-model reasoning.
30. Oct 12, 2026 · Research paper · 1 source
    VAttnPO: Visual Attention-Guided Policy Optimization
    We introduce VAttnPO (Visual Attention-Guided Policy Optimization), which uses answer-conditioned visual attention to refine token-level credit while preserving the verifier-derived trajectory-mean advantage.

Often appears with