Paper page - LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents
… The ledger is also used to check state-dependent policy constraints before environment-changing tool calls are executed, blocking policy violations. …
Search
Paper page - Policy and World Modeling Co-Training for Language Agents
Papers arxiv:2606.02388 Policy and World Modeling Co-Training for Language Agents Published on Jun 1 Submitted by Ning Lu on Jun 2 HKUST Authors: Ning Lu , , , , , , , , , , , Abstract PaW is a co-training framework that combines policy learning and world modeling using on-policy reinforcement lear… …
Paper page - LaWAM: Latent World Action Models for Efficient Dynamics-Aware Robot Policies
…but often lack explicit foresight into how robot actions change the scene. World-Action Models (WAMs) address this limitation by conditioning policies on predicted futures, yet existing approaches typically rely on computationally…
Paper page - AURA: Action-Gated Memory for Robot Policies at Constant VRAM
… They need to remember what changes the next action. …
Paper page - A Gradient Perspective on RLVR Stability and Winner Advantage Policy Optimization
… The following papers were recommended by the Semantic Scholar API ASymPO: Asymmetric-Scale Policy Optimization for Asynchronous LLM Post-Training Without Behavior Information 2026 Rethinking Importance Sampling in LLM Policy Optimization: A Cumulative Token Perspective 2026 Momentum for Reasoning: … …
Paper page - The Extrapolation Cliff in On-Policy Distillation of Near-Deterministic Structured Outputs
Papers arxiv:2605.08737 The Extrapolation Cliff in On-Policy Distillation of Near-Deterministic Structured Outputs Published on May 9 Submitted by XinLi on May 14 Nanyang Technological University Authors: Xin Li , , , , Abstract On-policy distillation with reward extrapolation exhibits a safety thr… …
Paper page - PoLAR: Factorizing Extent and Mode in Latent Actions for Robot Policy Learning
Papers arxiv:2606.21139 PoLAR: Factorizing Extent and Mode in Latent Actions for Robot Policy Learning Published on Jun 19 Submitted by Youngjoon Jeong on Jun 23 Seoul National University Authors: Youngjoon Jeong , , , , Abstract PoLAR introduces a geometrically structured latent action representat… …
Paper page - OPSD-V: On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators
… Generated by Qwen/Qwen2.5-Coder-32B-Instruct We propose OPSD-V, an on-policy self-distillation paradigm for post-training few-step autoregressive AR video diffusion models. …
Paper page - Dense Supervision, Sparse Updates: On the Sparsity and Geometry of On-Policy Distillation
… Generated by Qwen/Qwen2.5-Coder-32B-Instruct On-policy distillation OPD has recently become a prominent post-training recipe as it combines two desirable ingredients: on-policy student trajectories and dense teacher supervision , yet how this hybrid changes a model's parameters remains unclear. …
Unlocking Agentic RL Training for GPT-OSS: A Practical Retrospective
… Also, the fix only applies when on policy=True. How do you handle cases where the policy has drifted? Is there a threshold or mechanism to detect when to switch to off-policy mode? …