Proximity-Based Multi-Turn Optimization: Practical Credit Assignment for LLM Agent Training

TL;DR

ProxMO is a practical framework for multi-turn LLM agent training that improves credit assignment by dynamically adapting to task difficulty and semantic context, leading to better performance with minimal computational overhead.

Contribution

ProxMO introduces success-rate-aware modulation and proximity-based soft aggregation, enhancing credit assignment in real-world multi-turn LLM training scenarios.

Findings

Significant performance improvements on ALFWorld and WebShop benchmarks.

Mechanisms are lightweight and compatible with existing frameworks.

Ablation studies confirm the effectiveness of each component.

Abstract

Multi-turn LLM agents are becoming pivotal to production systems, spanning customer service automation, e-commerce assistance, and interactive task management, where accurately distinguishing high-value informative signals from stochastic noise is critical for sample-efficient training. In real-world scenarios, a failure in a trivial task may reflect random instability, whereas success in a high-difficulty task signifies a genuine capability breakthrough. Yet, existing group-based policy optimization methods rigidly rely on statistical deviation within discrete batches, frequently misallocating credit when task difficulty fluctuates. To address this issue, we propose Proximity-based Multi-turn Optimization (ProxMO), a practical and robust framework engineered specifically for the constraints of real-world deployment. ProxMO integrates global context via two lightweight mechanisms: success-rate-aware modulation dynamically adapts gradient intensity based on episode-level difficulty, while proximity-based soft aggregation derives baselines through continuous semantic weighting at the step level. Extensive evaluations on ALFWorld and WebShop benchmarks demonstrate that ProxMO yields substantial performance gains over existing baselines with negligible computational cost. Ablation studies further validate the independent and synergistic efficacy of both mechanisms. Crucially, ProxMO offers plug-and-play compatibility with standard GRPO frameworks, facilitating immediate, low-friction adoption in existing industrial training pipelines. Our implementation is available at: \href{https://anonymous.4open.science/r/proxmo-B7E7/README.md}{https://anonymous.4open.science/r/proxmo}.