Metraipolis › archive › dc_694603faa560

Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment — arXiv:2610.07023

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.07023
License (as archived)
CC0-1.0
Archived
2026-10-07T11:18:17.220Z
SHA-256 of the archived text
a088cc9004860d66d64a5215848d02f81cefff071e022f15d37a8aecff1887d2
Stable URL of this page
https://metraipolis.com/documento/dc_694603faa560

How to cite

Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment — arXiv:2610.07023 — arXiv (en), https://arxiv.org/abs/2610.07023 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_694603faa560 (Metraipolis archive dc_694603faa560, sha256 a088cc9004860d66d64a5215848d02f81cefff071e022f15d37a8aecff1887d2)
BibTeX
@misc{dc694603faa560,
  title = {Beyond Refusal Patterns: Safe-Role Internalization for Robust and Generalizable LLM Safety Alignment — arXiv:2610.07023},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.07023},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_694603faa560; sha256 a088cc9004860d66d64a5215848d02f81cefff071e022f15d37a8aecff1887d2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07023. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Language Models (LLMs) have achieved remarkable capabilities but remain vulnerable to jailbreak attacks that elicit harmful or unsafe outputs. Existing safety alignment approaches, including Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF), often require substantial attack-specific supervision and computational resources, while remaining susceptible to shallow safety alignment and over-refusal. To address these challenges, we introduce SSRFT(Supervised Safe-Role Fine-Tuning), the first framework that reformulates safety alignment as the internalization of a predefined safe role. SSRFT constructs a Safe-Role Question-Answer (SRQA) dataset from psychometric questions, limited jailbreak prompts, and a safe-role description. Role-consistent responses are synthesized, validated, and expanded into diverse scenarios, enabling models to internalize safety-oriented values and principles rather than explicit refusal patterns. Experiments across multiple Base and Instruct models show that SSRFT achieves more robust and generalizable safety alignment than standard SFT. SSRFT shows substantially greater robustness to prefilling attacks and better generalization to unseen jailbreak domains, while reducing over-refusal on benign queries and preserving the model's general capabilities. These results establish safe-role internalization as an effective alternative to refusal-centric safety alignment. Warning: This paper contains examples of harmful and toxic language.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2610.07023 — 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 2026-10-07T11:15:24.469Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).