Metraipolis › archive › dc_684b458433cc

HiRoute: Hierarchical Routed Prompt Tuning for Safety Alignment of Large Language Models — arXiv:2608.12821

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.12821
License (as archived)
CC0-1.0
Archived
2026-10-19T04:27:26.480Z
SHA-256 of the archived text
40b5b2e59f4a651c8b7bf334accc3f840285877cc97121fe5158fa54581ffa3b
Stable URL of this page
https://metraipolis.com/documento/dc_684b458433cc

How to cite

HiRoute: Hierarchical Routed Prompt Tuning for Safety Alignment of Large Language Models — arXiv:2608.12821 — arXiv (en), https://arxiv.org/abs/2608.12821 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_684b458433cc (Metraipolis archive dc_684b458433cc, sha256 40b5b2e59f4a651c8b7bf334accc3f840285877cc97121fe5158fa54581ffa3b)
BibTeX
@misc{dc684b458433cc,
  title = {HiRoute: Hierarchical Routed Prompt Tuning for Safety Alignment of Large Language Models — arXiv:2608.12821},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.12821},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_684b458433cc; sha256 40b5b2e59f4a651c8b7bf334accc3f840285877cc97121fe5158fa54581ffa3b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12821. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models (LLMs) remain vulnerable to harmful requests and jailbreak attacks. Parameter-efficient safety alignment methods based on prompt tuning typically rely on a single global prompt or externally selected prompt modules. Such static designs struggle to maintain a cross-category safety boundary while generating constructive responses tailored to specific risks and avoiding over-refusal of benign inputs. To address these limitations, we propose HiRoute, an input-adaptive hierarchical prompt-tuning framework that separates category-agnostic safety control from category-specific response guidance. HiRoute first trains a lightweight hierarchical router on representations extracted from a frozen LLM to jointly detect harmful intent and predict multi-label risk scores. It then freezes both the backbone model and the router and uses preference optimization with alternating gradient updates to learn a shared coarse-grained prompt and a set of fine-grained prompt experts as continuous embeddings. At inference time, benign inputs bypass the safety branch, whereas risky inputs are processed using the shared prompt together with a router-weighted mixture of risk-specific prompt experts. Experiments across three instruction-tuned models show that HiRoute achieves high safety rates across multiple safety benchmarks while preserving safe-response helpfulness, reducing over-refusal, and maintaining competitive performance on general-purpose tasks.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.12821 — 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 2026-10-19T04:20:35.828Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).