arXiv AI October 7, 2026 By Pratyay Dutta, Kowshik Thopalli, Vivek Narayanaswamy

COMPASS: Finding Where Reasoning Lives in Language Models

The paper introduces COMPASS, an inference-time steering technique that enhances large language model performance on mathematical reasoning tasks. By using the correctness of a model’s direct answer attempts as a latent signal, COMPASS identifies a small subset of attention heads whose activations can be steered to elicit reasoning. Experiments across three model families and multiple math benchmarks show that COMPASS improves GSM8K accuracy by an average of 16 percentage points and achieves near Chain‑of‑Thought performance while generating 20–70% fewer tokens.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv AI

Aug 5

Soft Guidance Starts to Outperform CoT Prompting as LLMs Improve

arXiv:2608. 03550v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting remains the standard baseline for evaluating models' reasoning abilities.

By Denys Pushkin, Albert Q. Jiang, Aryo Lotfi, Colin Sandon, Emmanuel Abb\'e

llms

arXiv Computation and Language

Aug 25

Mechanistic Interpretability of Chain-of-Thought Reasoning via Sequential Activation Patching

The paper introduces a sequential activation patching framework to study how Chain-of-Thought (CoT) prompting influences large language models over multiple generated tokens. By tracking CoT-conditioned attention-head activations across token positions and aggregating them with Part-of-Speech guidance, the authors identify distributed head sets that jointly contribute to answer generation. Targeted zero-ablation experiments confirm that these heads are functionally important, affecting mechanisms such as reasoning-trajectory maintenance, answer anchoring, exemplar-target separation, and numerical generation.

By Murat Dura, Serkan \"Ozt\"urk, Selma Tekir

llms nlp benchmarks safety

arXiv AI

Jul 24

Post-Hoc Reasoning in Chain of Thought: Decoding and Steering Pre-Committed Answers

arXiv:2603. 01437v2 Announce Type: replace Abstract: As chain of thought (CoT) has become central to scaling reasoning capabilities in large language models (LLMs), it has also emerged as a promising tool for interpretability, suggesting the opportunity to understand model decisions through verbalized reasoning.

By Kyle Cox, Darius Kianersi, Adri\`a Garriga-Alonso

llms safety

arXiv Machine Learning

Aug 11

Rethinking Reasoning with MDLMs: Early Exits, Post-hoc Reasoning, and Beyond

arXiv:2510. 19990v2 Announce Type: replace Abstract: The reasoning paradigm, where language models reason before answering, has enabled breakthroughs on tasks such as mathematical problem-solving.

By Zachary Horvitz, Raghav Singhal, Hao Zou, Carles Domingo-Enrich, Zhou Yu, Rajesh Ranganath, Kathleen McKeown

llms diffusion reinforcement-learning fine-tuning

arXiv AI

Sep 15

Post-Reasoning: Improving the Performance of Non-Thinking Models at No Cost

arXiv:2605.06165v2 Announce Type: replace Abstract: As the widespread adoption of Large Language Models (LLMs) accelerates, token consumption from intermediate reasoning traces increasingly contribut...

By Richmond Sin Jing Xuan, Rishabh Bhardwaj, Soujanya Poria

llms efficiency benchmarks

arXiv AI

Jul 14

Valid $\ne$ Necessary: Diagnosing Latent Inefficiency in Chain-of-Thought

arXiv:2607. 11266v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting has significantly advanced the reasoning capabilities of Large Language Models (LLMs), yet it often incurs substantial computational costs due to over-reasoning: the generation of redundant, verbose, or irrelevant steps.

By Daeyeop Lee, Hwanjo Yu

llms efficiency benchmarks

One email a morning, machine-written

Related stories

Soft Guidance Starts to Outperform CoT Prompting as LLMs Improve

Mechanistic Interpretability of Chain-of-Thought Reasoning via Sequential Activation Patching

Post-Hoc Reasoning in Chain of Thought: Decoding and Steering Pre-Committed Answers

Rethinking Reasoning with MDLMs: Early Exits, Post-hoc Reasoning, and Beyond

Post-Reasoning: Improving the Performance of Non-Thinking Models at No Cost

Valid $\ne$ Necessary: Diagnosing Latent Inefficiency in Chain-of-Thought