Metraipolis › archive › dc_d4e3d3217a3f

ALTSTEER: Selective Safety Steering for Moving Beyond Hard Refusals to Constructive Alternatives — arXiv:2608.30197

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.30197
License (as archived)
CC0-1.0
Archived
2026-10-17T08:07:24.113Z
SHA-256 of the archived text
f4f603f836ced3e256a46f035d14c3dc950a2592fc4ab4f2493d81ee88a45191
Stable URL of this page
https://metraipolis.com/documento/dc_d4e3d3217a3f

How to cite

ALTSTEER: Selective Safety Steering for Moving Beyond Hard Refusals to Constructive Alternatives — arXiv:2608.30197 — arXiv (en), https://arxiv.org/abs/2608.30197 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d4e3d3217a3f (Metraipolis archive dc_d4e3d3217a3f, sha256 f4f603f836ced3e256a46f035d14c3dc950a2592fc4ab4f2493d81ee88a45191)
BibTeX
@misc{dcd4e3d3217a3f,
  title = {ALTSTEER: Selective Safety Steering for Moving Beyond Hard Refusals to Constructive Alternatives — arXiv:2608.30197},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.30197},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d4e3d3217a3f; sha256 f4f603f836ced3e256a46f035d14c3dc950a2592fc4ab4f2493d81ee88a45191},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30197. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Safety alignment is essential for deploying large language models, requiring systems to prevent harmful compliance while preserving helpfulness on benign requests. Activation steering offers a training-free inference-time approach to safety control, but effective safety steering requires addressing two coupled questions: when to intervene and how generation should be shaped after intervention. However, existing safety steering methods remain limited along both dimensions, as their triggering mechanisms can be unstable across domains and refusal-oriented steering often yields rigid refusals rather than constructive safe guidance. To address these limitations, we propose ALTSTEER, an inference-time framework that couples selective intervention with refusal-anchored constructive redirection within a single inference pass. ALTSTEER uses an internal refusal-relevant signal to decide when to steer, and applies staged steering to shift generation from refusal-oriented control toward constructive alternatives. Evaluations on Llama-3.1 and Qwen2.5 show that ALTSTEER preserves benign utility while improving constructive safe-completion behavior, especially on models that otherwise tend to produce short refusals for harmful requests.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Software Engineering (cs.SE).
Pagina del lavoro: https://arxiv.org/abs/2608.30197 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T07:47:52.784Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).