Metraipolis › archive › dc_33ba0f4dd3fc
Anchor-ECC: Local Integrity Checking for Watermarked LLM Outputs via Error-Correcting Codes — arXiv:2609.38722
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.38722
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:14:20.877Z
- SHA-256 of the archived text
afe49efe330db1c3f7ee017213039ca0219424d04c900353d66f65e20f9e26b3- Stable URL of this page
https://metraipolis.com/documento/dc_33ba0f4dd3fc
How to cite
Anchor-ECC: Local Integrity Checking for Watermarked LLM Outputs via Error-Correcting Codes — arXiv:2609.38722 — arXiv (en), https://arxiv.org/abs/2609.38722 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_33ba0f4dd3fc (Metraipolis archive dc_33ba0f4dd3fc, sha256 afe49efe330db1c3f7ee017213039ca0219424d04c900353d66f65e20f9e26b3)
BibTeX
@misc{dc33ba0f4dd3fc,
title = {Anchor-ECC: Local Integrity Checking for Watermarked LLM Outputs via Error-Correcting Codes — arXiv:2609.38722},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.38722},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_33ba0f4dd3fc; sha256 afe49efe330db1c3f7ee017213039ca0219424d04c900353d66f65e20f9e26b3},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38722. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
LLM watermarking has become an effective approach to distinguishing AI-generated text from human-written text by embedding detectable patterns during generation. However, a small post-generation edit may change the meaning of the text without removing its overall watermark signal, creating a risk that the modified content is still attributed to the original model. We propose Anchor-ECC, which incorporates the error-correcting code (ECC) constraints and explicit boundary anchors into the watermark structure and pairs them with a dynamic-programming decoder to detect and localize post-generation edits. Across Qwen3-8B, Mistral-7B-Instruct-v0.3, and OPT-125M, the approximate-hard setting achieves about 99.7% block-level true positive rate (TPR) with at most 7.6% false alarm rate (FAR) for edit detection under mixed insertions, deletions, and substitutions, while preserving the distinction between watermarked outputs and unwatermarked text. Additional quality experiments identify lower-perplexity configurations that retain strong edit-detection performance. Together, these results extend LLM watermarking from source identification to local integrity verification while supporting configurable trade-offs between detection reliability and generation quality. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.CL). Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.38722 — 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 2026-10-07T12:44:51.942Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).