Metraipolis › archive › dc_04384d508af4
Nullify: Null-Space Activation Steering for Training-Free LLM Unlearning — arXiv:2610.10655
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.10655
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T05:53:52.728Z
- SHA-256 of the archived text
b860354373b1d337a5534249c35d01b0421e739f3e5af0d24c91e4e6bb3aaf3f- Stable URL of this page
https://metraipolis.com/documento/dc_04384d508af4
How to cite
Nullify: Null-Space Activation Steering for Training-Free LLM Unlearning — arXiv:2610.10655 — arXiv (en), https://arxiv.org/abs/2610.10655 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_04384d508af4 (Metraipolis archive dc_04384d508af4, sha256 b860354373b1d337a5534249c35d01b0421e739f3e5af0d24c91e4e6bb3aaf3f)
BibTeX
@misc{dc04384d508af4,
title = {Nullify: Null-Space Activation Steering for Training-Free LLM Unlearning — arXiv:2610.10655},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.10655},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_04384d508af4; sha256 b860354373b1d337a5534249c35d01b0421e739f3e5af0d24c91e4e6bb3aaf3f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.10655. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large Language Models (LLMs) inevitably internalize substantial amounts of sensitive or private information during pre-training, while LLM unlearning aims to selectively erase specific knowledge to prevent privacy leakage with minimal loss of model utility. However, existing methods struggle to balance forget quality with utility, and typically incur substantial computational costs due to parameter fine-tuning. To address this, we propose Nullify, a training-free, non-destructive activation steering method for LLM unlearning. Nullify employs steering vectors during inference to redirect privacy-related activations away from their memorized answers, while satisfying a null-space constraint that leaves retained-query activations essentially unaffected to maintain utility. Evaluations on TOFU and MUSE show that Nullify matches or surpasses established baselines in forget quality while achieving near-lossless preservation of model utility. By avoiding weight updates entirely, Nullify serves as an efficient, plug-and-play inference-time intervention framework. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2610.10655 — 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 2026-10-09T05:47:01.302Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).