@HuggingPapers: When should LLMs update, preserve, or ignore information? Contextual Belief Management is what long-horizon reasoning w…
Summary
Introduces BeliefTrack, a method for contextual belief management in LLMs, reducing reasoning failures by over 70%.
View Cached Full Text
Cached at: 05/31/26, 04:58 AM
When should LLMs update, preserve, or ignore information?
Contextual Belief Management is what long-horizon reasoning was missing. We introduce BeliefTrack—and show that optimizing belief states cuts reasoning failures by over 70%. https://t.co/7gwuNLNd1t
Similar Articles
When Should Models Change Their Minds? Contextual Belief Management in Large Language Models
This paper introduces Contextual Belief Management (CBM) for LLMs to handle long-term information, proposes the BeliefTrack benchmark for evaluation, and demonstrates that reinforcement learning and representation-level steering significantly reduce belief management failures.
Towards a Belief-Based World Model for LLM Agents
This paper introduces Belief-Based World Models (BB-WMs) to enhance LLM agents' decision-making under partial observability by providing direct access to beliefs about uncertain states, showing improved task performance.
Parallel LLM Reasoning for Bias-Resilient, Robust Conceptual Abstraction
This paper proposes a framework for parallel chunk-level processing of long documents with LLMs to reduce cumulative bias and improve evidence traceability, achieving significant reductions in omission errors and unsupported claims.
Adaptive Triggering for Bias Correction in LLM Reasoning
This paper introduces an adaptive triggering framework for bias correction in LLM chain-of-thought reasoning, using online change-point detection to optimize intervention timing with white-box and black-box signals, improving accuracy and reducing interventions.
LLMs know when they are wrong. I made a fix relating to Anthropic's new "global workspace" paper [R]
The author presents a method to make LLMs verbalize calibrated confidence by using a linear probe on mid-layer states and a small trained bridge to confidence logits, requiring only 200 labeled examples and no weight modification. This is linked to Anthropic's global workspace paper explaining the know-say gap.