| title | Capable yet Parsimonious: Extracting and Characterizing Hidden Chain-of-Thought in Frontier Models |
|---|---|
| authors | Xiaoyu Luo, Tao Ren, Wenrui Yu, Xiao Li, Qiongxiu Li, Johannes Bjerva |
| date | 2026-09-22 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2609.26637 |
| affiliation | Sea-Fill |
| upvotes | 24 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | The rapid capability gains of frontier language models are widely attributed to improved reasoning abilities, yet this cannot be verified as raw CoT traces in closed-source systems are hidden. By registering a simple custom tool through a standard API feature, we induce frontier models to externalize intermediate reasoning. Because these traces may reflect post-hoc rationalization rather than genuine reasoning, we first evaluate against native CoT on open-source models and extend to closed-source frontier models including GPT-6 Astra. We find that the extracted reasoning matches native reasoning performance and substantially outperforms no-reasoning baselines, across competition mathematics, science, and code generation. We then characterize how frontier models structure their intermediate reasoning. Across token efficiency, reasoning-step types, and induced reasoning trees, we identify systematic differences in how models externalize, compress, and organize reasoning. We find that Astra exhibits token-efficient directed reasoning, selecting a correct trajectory earlier, while resolving elementary steps internally and externalizing only crucial reasoning. These findings provide a behavioral lens on frontier-model reasoning beyond benchmark scores. |
cite
@misc{luo2026capable,
title={Capable yet Parsimonious: Extracting and Characterizing Hidden Chain-of-Thought in Frontier Models},
author={Xiaoyu Luo and Tao Ren and Wenrui Yu and Xiao Li and Qiongxiu Li and Johannes Bjerva},
year={2026},
eprint={2609.26637},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2609.26637}
}