Granularity-Adaptive Credit Assignment for Long-Horizon LLM Agent Reinforcement Learning

• URL: http://arxiv.org/abs/2609.12424v1
• Date: Fri, 11 Sep 2026 04:19:44 GMT
• Title: Granularity-Adaptive Credit Assignment for Long-Horizon LLM Agent Reinforcement Learning
• Abstract summary: We argue that the right resolution is state-dependent, and propose GACA, a critic-free estimator whose granularity follows an uncertainty-based criticality proxy.<n>GACA scores every step by the negative log-likelihood its own rollout already records, then blends the two advantages with a per-step weight that grows with that score.<n>On ALFWorld and WebShop, GACA improves task success over GRPO and GiGPO at both 1.5B and 7B scales.
• Score: 7.885980895039013
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Reinforcement learning is now the standard way to train large language model agents on long-horizon tasks, where dozens of interdependent actions precede a single sparse reward. Critic-free, group-relative methods such as GRPO suit this regime, but they broadcast one trajectory-level scalar to every step and cannot say which decision drove the outcome. GiGPO recovers a step-level signal by grouping time steps that share an anchor state, yet it merges the step- and episode-level estimates under one fixed weight, spending the same resolution on a pivotal branching decision as on a routine, near-deterministic transition. We argue that the right resolution is state-dependent, and propose GACA, a critic-free estimator whose granularity follows an uncertainty-based criticality proxy. GACA scores every step by the negative log-likelihood its own rollout already records, then blends the two advantages with a per-step weight that grows with that score, so the gradient places more weight on the fine-grained signal at above-average NLL and on the episode-level signal below it. We derive an exact risk decomposition for the implemented mixture and show that sufficiently small modulation improves on fixed mixing under positive directional alignment. A separate conditional result bounds local action-value variation using expected NLL, while an error-projection analysis characterizes when mixing adds value beyond scalar uncertainty reweighting. On ALFWorld and WebShop, GACA improves task success over GRPO and GiGPO at both 1.5B and 7B scales.

Related papers

• One Step, One Lead: Mitigating Higher-Order Interference in Multi-Domain Reinforcement Learning via Cross-Step Control [62.82372407840088]
  We show that same-point domain gradients may remain nearly even when consecutive realized updates partially reverse one another in output space.<n>We propose OSOL, which designates a focus domain at each iteration, to rank token-level rebound risk.<n>OSOL reaches a domain-macro average of 0.4822, improving by 5.7% over the strongest compared baseline, without explicit higher-order differentiation.
  arXiv  Detail & Related papers  (2026-09-06T08:32:59Z)
• The Dark Room in the Reward Channel: Dense Prediction Rewards Collapse GRPO-Trained LLM Agents -- and What Actually Works [3.040305634723913]
  We show that under group-normalized RL, a potential-based prediction reward drives every run into a degenerate absorbing state.<n>A single-episode ablation ablations the cause -- removing only GRPO's normalization turns the same reward from catastrophic into baseline parity.<n>Our central insight generalizes this: what z-scoring amplifies is a dense signal's within-group variance while all-fail groups dominate.
  arXiv  Detail & Related papers  (2026-07-23T12:50:18Z)
• Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning [85.02566758103008]
  Group-Graph Policy Optimization (G2PO) is a novel group-based Reinforcement Learning algorithm tailored for multi-turn agentic tasks.<n>G2PO substantially outperforms state-of-the-art prompt-based and RL baselines, achieving remarkable success rate improvements of up to 22.2% over GRPO.
  arXiv  Detail & Related papers  (2026-06-22T08:12:47Z)
• RL2ML: Finite-Rollout Surrogate Objectives from Reinforcement Learning to Maximum Likelihood [0.0]
  This paper develops RL2ML, a family of finite-rollout surrogate objectives with a closed-form, exactly unbiased gradient estimator.<n>We show that the best choice of surrogate objective is determined neither by proximity to maximum likelihood nor by the population-level weight alone.
  arXiv  Detail & Related papers  (2026-05-28T16:14:45Z)
• StepOPSD: Step-Aware Online Preference Distillation for Agent Reinforcement Learning [6.365332042924078]
  We present StepOPSD, a post-rollout preference self-distillation framework that takes the agent step as the unit of credit redistribution.<n>StepOPSD decomposes trajectories into action-centered step segments, rescoring them under hindsight-enriched teacher contexts.<n>Results reveal a consistent two-knob law: smaller _clip acts as a broadly stabilizing local trust region, whereas the optimal global mixing strength _mix remains task-dependent.
  arXiv  Detail & Related papers  (2026-05-26T15:07:03Z)
• Clipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary Signals [83.0127582612634]
  Near-boundary Rescue (NSR) is a minimal, plug-and-play modification that retains out-of-bound tokens to recover lost signals.<n>NSR substantially improves training stability and delivers consistent gains over strong baselines such as DAPO and GSPO.
  arXiv  Detail & Related papers  (2026-05-21T16:45:31Z)
• Holder Policy Optimisation [26.521180498291717]
  textbfHlderPO is a generalised policy optimisation framework.<n>It unifies token-level probability aggregation via the Hlder mean.<n>It achieves a state-of-the-art average accuracy of $54.9%$ across multiple mathematical benchmarks.
  arXiv  Detail & Related papers  (2026-05-12T12:45:03Z)
• AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models [60.04879435087352]
  Vision-language-action (VLA) policies generate actions within a single unified space.<n>We propose AnchorRefine, a hierarchical framework that factorizes VLA action modeling into trajectory anchor and residual refinement.<n> Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones.
  arXiv  Detail & Related papers  (2026-04-20T04:25:24Z)
• Improving Search Agent with One Line of Code [68.58667107354253]
  Tool-based Agentic Reinforcement Learning (TARL) has emerged as a promising paradigm for training search agents to interact with external tools.<n>We propose textbfSearch textbfAgent textbfPolicy textbfOptimization (textbfSAPO), which stabilizes training via a conditional token-level KL constraint.
  arXiv  Detail & Related papers  (2026-03-10T04:07:39Z)
• Anchoring Values in Temporal and Group Dimensions for Flow Matching Model Alignment [61.80228667422234]
  VGPO redefines value estimation across both temporal and group dimensions.<n>It transforms the sparse terminal reward into dense, process-aware value estimates.<n>It replaces standard group normalization with a novel process enhanced by absolute values to maintain a stable optimization signal.
  arXiv  Detail & Related papers  (2025-12-13T16:31:26Z)

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.