# Better Accuracies, Worse Reasoning: A Step-Level Audit of Medical Chain-of-Thought Distillation

arXiv: https://arxiv.org/abs/2605.28301 · HTML: https://mlsanity.com/paper/2605.28301/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Clinical and biomedical applications, one of 17 directions in LLM reasoning.
- Builds on 18 of its 49 references that are in the map.
- First posted on arXiv 27 May 2026.

## Placed in

- LLM reasoning / [Clinical and biomedical applications](https://mlsanity.com/lr/clinical-and-biomedical-applications/)
- Knowledge distillation / [Knowledge distillation for LLMs](https://mlsanity.com/distill/knowledge-distillation-for-llms/)

## Builds on (most cited first, 18 of 18)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [Solving math word problems with process- and outcome-based feedback](https://mlsanity.com/paper/2211.14275/) (2022)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [Discovering Latent Knowledge in Language Models Without Supervision](https://mlsanity.com/paper/2212.03827/) (2022)
- [Teaching Models to Express Their Uncertainty in Words](https://mlsanity.com/paper/2205.14334/) (2022)
- [Measuring Faithfulness in Chain-of-Thought Reasoning](https://mlsanity.com/paper/2307.13702/) (2023)
- [Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation](https://mlsanity.com/paper/2503.11926/) (2025)
- [Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations](https://mlsanity.com/paper/1703.03717/) (2017)
- [Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety](https://mlsanity.com/paper/2507.11473/) (2025)
- [MedReason: Eliciting Factual Medical Reasoning Steps in LLMs via Knowledge Graphs](https://mlsanity.com/paper/2504.00993/) (2025)
- [Common 7B Language Models Already Possess Strong Math Capabilities](https://mlsanity.com/paper/2403.04706/) (2024)
- [Are Reasoning Models More Prone to Hallucination?](https://mlsanity.com/paper/2505.23646/) (2025)
- [Why Chain of Thought Fails in Clinical Text Understanding](https://mlsanity.com/paper/2509.21933/) (2025)

## Built on by (most cited first, 2 of 2)

- [Hindsight-Guided Rationale Distillation for Rare Disease Diagnosis](https://mlsanity.com/paper/2610.03176/) (2026)
- [MedEvo: Medical Reasoning Evolution through Knowledge Navigation](https://mlsanity.com/paper/2610.17690/) (2026)

## Abstract (arXiv)

Chain-of-thought (CoT) distillation trains a smaller model to imitate a teacher's reasoning trace, but it is typically evaluated by final-answer metrics including accuracy. We ask whether gains in answer quality are accompanied by improvements in the trace. In medical QA, where short answer options can leave a richer clinical justification under-specified, a Qwen3-8B student distilled from a DeepSeek-V3-family teacher improves on MedQA-USMLE answer metrics (SC@64 74.7% to 84.4%; expected calibration error (ECE) 0.096 to 0.034). Yet under a Kimi-K2.6 style-blind LLM-judge audit, its error rate over non-abstained steps rises from 30.6% to 50.3%. In this primary medical setting, answer quality and trace factuality move in opposite directions. This before--after pattern persists across evaluators, teacher strengths, student scales and families, medical benchmarks, and style, segmentation, and answer-correctness controls. A 150-step blinded audit by a clinical expert reproduces the same ordering. Boundary checks narrow the scope of the claim: the risk appears when a compact answer under-constrains the rationale and a capable student can imitate expert-like form without reliably grounding each local claim. Standard answer metrics and aggregate hedging rates do not reveal the shift. When such traces are released or reused, answer-level metrics alone are insufficient.