ReasoningLens: Hierarchical Visualization and Diagnostic Auditing for Large Reasoning Models

2026-06-22

AI Summary

This work addresses the opacity of lengthy reasoning chains generated by large reasoning models, wherein critical logical steps are often obscured. To enhance interpretability, the authors propose the first open-source framework supporting hierarchical interactive visualization, which decouples unstructured reasoning traces into high-level strategies and low-level executions. The framework incorporates an intelligent auditing agent augmented with tool-based verification mechanisms to automatically detect errors, uncover reasoning blind spots, and construct systematic reasoning profiles. This approach substantially improves the explainability, debuggability, and optimization efficiency of model reasoning processes.

Abstract

The emergence of Large Reasoning Models has introduced exceptionally long Chain-of-Thought traces, creating a transparency burden where critical logic is often buried under massive procedural text. To address this, we present ReasoningLens, an open-source framework designed for the hierarchical visualization and diagnostic auditing of complex reasoning chains. ReasoningLens addresses information necropsy by: (1) structuring traces into interactive hierarchies that separate high-level strategy from low-level execution; (2) leveraging an agentic auditor for automated error detection and tool-augmented verification; and (3) synthesizing systemic reasoning profiles to reveal model-specific blind spots. By transforming unstructured walls of text into actionable insights, ReasoningLens provides a modular foundation for interpreting, debugging, and optimizing the next generation of reasoning-centric AI.

Problem

• Large Reasoning Models
• Chain-of-Thought
• transparency
• reasoning trace
• diagnostic auditing

Innovation

• hierarchical visualization
• diagnostic auditing
• Chain-of-Thought
• agentic auditor
• reasoning profiles