资讯

Safety Side Effects of Preventative Persona Steering: An Empirical Audit

📅 2026-11-16 ⏱️ 约 5 分钟阅读 ✍️ AI导航编辑部 🔗 arxiv.org
Safety Side Effects of Preventative Persona Steering: An Empirical Audit
📝 内容摘要

Computer Science > Artificial Intelligence [Submitted on 13 Nov 2026] Title:Safety Side Effects of Preventative Pe

📌 核心要点

  • Computer Science > Artificial Intelligence
  • [Submitted on 13 Nov 2026]
  • Title:Safety Side Effects of Preventative Persona Steering: An Empirical Audit

Computer Science > Artificial Intelligence

[Submitted on 13 Nov 2026]

Title:Safety Side Effects of Preventative Persona Steering: An Empirical Audit

View PDF HTML (experimental)Abstract:High-level behaviors such as sycophancy and hallucination correspond to directions in a model's activation space (Chen et al., 2025). These directions, known as persona vectors, can suppress undesirable traits either at inference time, subtracting the vector during generation, or through preventative steering, adding the vector during fine-tuning, so the optimizer never needs to encode the trait in weights, then removing the vector at evaluation. Chen et al. (2025) recommend preventative steering because it preserves Massive Multitask Language Understanding (MMLU) accuracy, but validations of this recommendation have measured capability alone. We primarily evaluate results on preventative steering of sycophancy and hallucination vectors on three models, measuring three safety benchmarks alongside trait expression, coherence, and MMLU. Our findings show that preventative steering still changes refusal behavior, but by an amount that depends on the trait and does not establish that preventative persona steering is safety-neutral, and steering methods should be evaluated on several safety benchmarks.

Additional Features

References & Citations

Loading...

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)

Connected Papers (What is Connected Papers?)

Litmaps (What is Litmaps?)

scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub (What is DagsHub?)

Gotit.pub (What is GotitPub?)

Hugging Face (What is Huggingface?)

ScienceCast (What is ScienceCast?)

Demos

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)

CORE Recommender (What is CORE?)

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

来源arxiv.org· 本文为编辑整理,仅供参考