About this paper
Language-model agents trained with reinforcement learning can improve by learning to create and curate experience artifacts that help future rollouts, not just by optimizing actions from scalar rewards.
It is common to train Large Language Models (LLMs) for agentic tasks with Reinforcement Learning (RL) from scalar rewards. However, this throws away the fine-grained information about the environment that the agent experienced during the trial (e.g., error logs). To utilize this rich information, recent works wrap the agent in a harness that refines its own inputs through reflection or compaction and hands past experience to future rollouts, but the construction of the handed-off artifacts (e.g., a playbook) is itself not learned. We want the best of both worlds: gradient updates from a scalar reward, while also utilizing environment information to hand off experience. As a concrete…
Large Language ModelsReinforcement Learningscalar rewardsagentic tasksagent harnessreflectioncompactionrolloutsgradient updatesAgentic Context Engineering
- Upvotes
- 4
- Authors
- 4
- Published
- Nov 12, 2026
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.
Trending history
Experience-Handoff Reinforcement Learning for Recursive Self-Improvement was first seen on Hugging Face · Daily papers on Nov 14, 2026. Since then it has appeared in 16 checks on 3 different days. Its best position was #21 of 30, reached on Nov 16, 05:28 UTC. It is #21 right now, climbing from #29 to #21 over the last 6 checks.