Beyond a Single Direction: Chain-of-Thought Disrupts Simple Steering of Refusal
The paper discusses how chain-of-thought (CoT) in large reasoning models (LRMs) complicates the control of refusal mechanisms. It highlights that refusal is influenced by both the CoT and residual stream activations, making LRMs more resilient to simple activation-level interventions. The findings suggest that the CoT can independently carry compliance signals, indicating a need for more nuanced control strategies.
- ▪Refusal in LRMs is influenced by chain-of-thought (CoT) and residual stream activations.
- ▪Activation steering reverses refusal in only 39% of cases when CoT is fixed, but increases to 70% when CoT is removed.
- ▪A two-stage intervention can reverse refusal in 94% of cases, demonstrating the CoT's role in compliance.
arXiv cs.AI files mainly under ai research. We currently carry 1,128 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | arXiv cs.AI |
| Canonical URL | https://arxiv.org/abs/2605.26772 |
| Publication time | Wed, 27 May 2026 00:00:00 -0400 |
| Retrieval time | 2026-05-27T04:07:56.398Z |
| Last seen | 2026-05-27T04:07:56.398Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | bP7T-gsgeclR |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.26772 (cs) [Submitted on 26 May 2026] Title:Beyond a Single Direction: Chain-of-Thought Disrupts Simple Steering of Refusal Authors:Kia-Jüng Yang, Dominik Meier, Jiachen Zhao, Terry Ruas, Bela Gipp View a PDF of the paper titled Beyond a Single Direction: Chain-of-Thought Disrupts Simple Steering of Refusal, by Kia-J\"ung Yang and 4 other authors View PDF HTML (experimental) Abstract:Large reasoning models (LRMs) generate chain-of-thought (CoT) traces before producing final outputs, introducing a dynamic internal state that may complicate control mechanisms such as refusal. Unlike instruction-tuned LLMs, where refusal is mediated by a single directional subspace, refusal in large reasoning models (LRMs) additionally depends on the CoT.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
