让下一篇重要论文，不再错过。

原文摘要

Large language models increasingly act as autonomous agents in long-horizon interactions, learning from final outcomes and environmental feedback. However, optimizing from such interaction experience remains challenging, as feedback is heterogeneous and rarely carries explicit reward information. Existing post-training approaches rely on outcome-driven methods, such as reinforcement learning with verifiable rewards (RLVR), which primarily exploit final success signals while leaving interaction experience and feedback underutilized. In sparse-reward, long-horizon tasks, this often results in distribution sharpening: the policy reinforces a narrow set of already-successful behaviors, without substantially improving the feedback-grounded agency needed for broader problem-solving capacity (e.g., Pass@k). We propose Counterfactual Distillation, a framework that converts exploration-derived experience into trainable policy supervision. Specifically, we organize exploration as a tree-structured search process, where the agent reflects on past decision points and makes new experience-guided decisions to form alternative branches. These corrections are then distilled as counterfactual targets: we train the model to predict the revised action from the original history without explicitly providing the experience, thereby introducing out-of-distribution behavior patterns and expanding exploration capacity. Across diverse interactive coding and agentic tasks, our method outperforms outcome-driven baselines such as GRPO and experience-based methods such as Early Experience, with gains of up to 14 percentage points over the base model on Pass@128. When interleaved with reinforcement learning updates, it further raises the performance ceiling, improving Pass@1 on Sokoban by 8 percentage points over extended GRPO. Our code is available at https://github.com/motigrez/Counterfactual-Distillation.