Refusal Lives Downstream of Persona in Chat Models
Researchers have found that refusal and persona traits in chat models are interconnected, with a compliant persona suppressing refusal. This discovery was made by studying instruction-tuned chat models, including Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct. The study's findings have implications for the development of more effective and compliant chat models.
- ▪A compliant persona direction and a refusal direction were extracted from the models and intervened on.
- ▪Compliant persona steering suppresses refusal, with the refusal rate falling from 97% to 2% in Llama.
- ▪Refusal is gated at the late-layer expression stage, downstream of where it is computed.
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.org |
| Canonical URL | https://arxiv.org/abs/2606.26161 |
| Publication time | Fri, 26 Jun 2026 00:00:00 -0400 |
| Retrieval time | 2026-06-26T05:20:40.857Z |
| Last seen | 2026-06-26T05:20:40.857Z |
| 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 | G2Z6_8AP0ShO |
| 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:2606.26161 (cs) [Submitted on 24 Jun 2026] Title:Refusal Lives Downstream of Persona in Chat Models Authors:Viola Zhong, Qirui Li View a PDF of the paper titled Refusal Lives Downstream of Persona in Chat Models, by Viola Zhong and 1 other authors View PDF HTML (experimental) Abstract:Linear directions in activation space have been identified for both refusal and persona traits in instruction-tuned chat models, but the two have been studied as separate mechanisms. We show they interact: a compliant persona gates refusal. In Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct, we extract a compliant model-persona direction and a refusal direction and intervene on both.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
