ABBEL: LLM Agents Acting through Belief Bottlenecks Expressed in Language

📝 Original Info

• Title: ABBEL: LLM Agents Acting through Belief Bottlenecks Expressed in Language
• ArXiv ID: 2512.20111
• Date: 2025-12-23
• Authors: Aly Lidayan, Jakob Bjorner, Satvik Golechha, Kartik Goyal, Alane Suhr

📝 Abstract

As the length of sequential decision-making tasks increases, it becomes computationally impractical to keep full interaction histories in context. We introduce a general framework for LLM agents to maintain concise contexts through multi-step interaction: Acting through Belief Bottlenecks Expressed in Language (ABBEL), and methods to further improve ABBEL agents with RL post-training. ABBEL replaces long multi-step interaction history by a belief state, i.e., a natural language summary of what has been discovered about task-relevant unknowns. Under ABBEL, at each step the agent first updates a prior belief with the most recent observation from the environment to form a posterior belief, then uses only the posterior to select an action. We systematically evaluate frontier models under ABBEL across six diverse multi-step environments, finding that ABBEL supports generating interpretable beliefs while maintaining near-constant memory use over interaction steps. However, bottleneck approaches are generally prone to error propagation, which we observe causing inferior performance when compared to the full context setting due to errors in belief updating. Therefore, we train LLMs to generate and act on beliefs withi...

📄 Full Content

...(본문 내용이 길어 생략되었습니다. 사이트에서 전문을 확인해 주세요.)