CLEAR: Context Augmentation from Contrastive Learning of Experience via Agentic Reflection

TL;DR

CLEAR introduces a generative context augmentation framework that uses contrastive learning and agentic reflection to produce task-specific knowledge, improving large language model agent performance.

Contribution

It proposes a novel contrastive learning-based method for context augmentation that outperforms retrieval-based approaches in LLM agents.

Findings

Improves task completion rate from 72.62% to 81.15% on AppWorld.

Increases average reward from 0.68 to 0.74 on WebShop.

Outperforms strong baseline methods across benchmarks.

Abstract

Large language model agents rely on effective model context to obtain task-relevant information for decision-making. Many existing context engineering approaches primarily rely on the context generated from the past experience and retrieval mechanisms that reuse these context. However, retrieved context from past tasks must be adapted by the execution agent to fit new situations, placing additional reasoning burden on the underlying LLM. To address this limitation, we propose a generative context augmentation framework using Contrastive Learning of Experience via Agentic Reflection (CLEAR). CLEAR first employs a reflection agent to perform contrastive analysis over past execution trajectories and summarize useful context for each observed task. These summaries are then used as supervised fine-tuning data to train a context augmentation model (CAM). Then we further optimize CAM using reinforcement learning, where the reward signal is obtained by running the task execution agent. By learning to generate task-specific knowledge rather than retrieve knowledge from the past, CAM produces context that is better tailored to the current task. We conduct comprehensive evaluations on the AppWorld and WebShop benchmarks. Experimental results show that CLEAR consistently outperforms strong baselines. It improves task completion rate from 72.62% to 81.15% on AppWorld test set and averaged reward from 0.68 to 0.74 on a subset of WebShop, compared with baseline agent. Our code is publicly available at https://github.com/awslabs/CLEAR.