Metraipolis › archive › dc_eae26e393d2f
Machine Unlearning for Large Language Models: Foundations, Advances, and Agentic Extensions — arXiv:2609.30909
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.30909
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T02:09:20.138Z
- SHA-256 of the archived text
104af69a893be9468a70fff0976c9de52c40f7bc8af0bde9dc453d3e5771d3a6- Stable URL of this page
https://metraipolis.com/documento/dc_eae26e393d2f
How to cite
Machine Unlearning for Large Language Models: Foundations, Advances, and Agentic Extensions — arXiv:2609.30909 — arXiv (en), https://arxiv.org/abs/2609.30909 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_eae26e393d2f (Metraipolis archive dc_eae26e393d2f, sha256 104af69a893be9468a70fff0976c9de52c40f7bc8af0bde9dc453d3e5771d3a6)
BibTeX
@misc{dceae26e393d2f,
title = {Machine Unlearning for Large Language Models: Foundations, Advances, and Agentic Extensions — arXiv:2609.30909},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.30909},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_eae26e393d2f; sha256 104af69a893be9468a70fff0976c9de52c40f7bc8af0bde9dc453d3e5771d3a6},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30909. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Machine unlearning aims to remove target influence while preserving other capabilities. This survey compares methods, benchmarks, and evidence across large language models and systems using retrieval, memory, tools, and interacting agents. A five-layer framework connects removal requests, system boundaries, target locations, interventions, and supported claims. A seven-stage lifecycle and six evidence dimensions guide comparison. The review shows that target construction, retained data, and recovery tests affect reported outcomes. Evidence from model evaluations remains insufficient to establish removal across external state and subsequent updates, motivating evaluation that tracks dependencies and tests whether target influence returns. Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-28 in cs.CR). Categorie arXiv: Cryptography and Security (cs.CR). Pagina del lavoro: https://arxiv.org/abs/2609.30909 — 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 2026-10-08T01:42:25.534Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).