Reasoning Trace Analysis

Follow this topic Ask about this topic

Reasoning trace analysis examines how language models and agents develop, revise, and use intermediate steps—not just whether their final answers are correct. Researchers study traces with step-level annotations and benchmarks, statistical analyses of text and hidden-state trajectories, and causal tools such as activation patching; they use these methods to detect errors, shortcuts, harmful behavior, and opportunities to stop or compress reasoning. A central tension is that a plausible trace may be invalid or unrelated to the answer, while internal signals can reveal useful information that the visible text omits. Current work therefore seeks evaluations that test trace validity and causal influence while balancing faithfulness, safety, and inference cost.

Latest papers

125 papers6 papers in the last four weeks, down 14% on the four weeks before. 0.0% of all new papers.

Bypassing the Rationale: Causal Auditing of Implicit Reasoning in Language Models
Feb 3, 2026Anish Sathyanarayanan, Aditya Nagarsekar, Aarush RathoreCoT Faithfulness Faithfulness of Language Model Explanations
BITS Pilani, K. K. Birla Goa Campus
Step-Tagging: Toward controlling the generation of Language Reasoning Models through step monitoring
Dec 16, 2025Yannis Belkhiter, Seshu Tirupathi, Giulio Zizzo +1Reasoning Trace Analysis Efficient Language Model Reasoning
1IBM Research Europe, Dublin, Ireland · 2Trinity College Dublin, Ireland · 3ADAPT Research Centre, Dublin, Ireland
Superficial Reflection or Genuine Thought? A Fine-Grained Cognitive Analysis of Large Reasoning Models
Nov 30, 2025Yuxiang Chen, Zuohan Wu, Ziwei Wang +6LLM Self-Correction Cognitive Modeling
University College London · AI Lab, the Yangtze River Delta, China · The Hong Kong University of Science and Technology (Guangzhou) +4
CoT-X: An Adaptive Framework for Cross-Model Chain-of-Thought Transfer and Optimization
Nov 7, 2025Ziqian Bi, Yinzhi Wang, Tianyang Wang +6Efficient Language Model Reasoning LLM Inference Acceleration
1Purdue University, USA · 2Baruch College, City University of New York, USA · 3The Ohio State University, USA +6
Base Models Know How to Reason, Thinking Models Learn When
Oct 8, 2025Constantin Venhoff, Iván Arcuschin, Philip Torr +2LLM Interpretability RL for Language Model Reasoning
University of Oxford, UK · MATS · Poseidon Research.

Feb 3, 2026J/K move · Enter open · S save

Bypassing the Rationale: Causal Auditing of Implicit Reasoning in Language Models

Anish Sathyanarayanan, Aditya Nagarsekar, Aarush Rathore

BITS Pilani, K. K. Birla Goa Campus

CoT Faithfulness Faithfulness of Language Model Explanations LLM Auditing Implicit Reasoning in Language Models Causal Interpretability Causal Interventions in Language Models+2

Open paper AskPDF arXiv

Chain-of-thought (CoT) prompting is widely used as a reasoning aid and is often treated as a transparency mechanism. Yet behavioral gains under CoT do not imply that the model's internal computation causally depends on the emitted reasoning text, i.e. models may produce fluent rationales while routing decision-critical computation through latent pathways. We introduce a causal, layerwise audit of CoT faithfulness based on activation patching. Our key metric, the CoT Mediation Index (CMI), isolates CoT-specific causal influence by comparing performance degradation from patching CoT-token hidden states against matched control patches. Across multiple model families (Phi, Qwen, DialoGPT) and scales, we find that CoT-specific influence is typically depth-localized into narrow ''reasoning windows,'' and we identify bypass regimes where CMI is near-zero despite plausible CoT text. We further observe that models tuned explicitly for reasoning tend to exhibit stronger and more structured mediation than larger untuned counterparts, while Mixture-of-Experts models show more distributed mediation consistent with routing-based computation. Overall, our results show that CoT faithfulness varies substantially across models and tasks and cannot be inferred from behavior alone, motivating causal, layerwise audits when using CoT as a transparency signal.

Latest papers

Bypassing the Rationale: Causal Auditing of Implicit Reasoning in Language Models

Step-Tagging: Toward controlling the generation of Language Reasoning Models through step monitoring

Superficial Reflection or Genuine Thought? A Fine-Grained Cognitive Analysis of Large Reasoning Models

CoT-X: An Adaptive Framework for Cross-Model Chain-of-Thought Transfer and Optimization

Base Models Know How to Reason, Thinking Models Learn When