When Steering Fails in Latent Reasoning: A Latent-to-Language Transition Gap
Activation steering has become a widely used approach for controlling language models during explicit chain-of-thought (CoT) reasoning, motivating its extension to latent CoT. However, we find that steering continuous thoughts produces substantially weaker effects on subsequent language generation than steering explicit CoT, even when the hidden representations are moved by comparable amounts. We first show that task information remains identifiable in continuous thoughts. Hence, we hypothesize a \textbf{latent-to-language transition gap}, in which an intervention effect in latent space fails to transfer to language generation. Two further results support this hypothesis: the output distr...
| Authors | Gaoxiang Huang, Lei Qi |
|---|---|
| Primary category | cs.CL |
| Published | 18 September 2026 |
| Last updated at source | 18 September 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.