LATIDIA · Ciberseguridad

RMCW: Una marca de agua de eliminación y eliminación basada en Reed: códigos de Muller para modelos de lenguaje

arXiv: 2610.02817v1Tipo de anuncio: nuevo Resumen: La marca de agua del modelo de lenguaje grande (LLM) proporciona un mecanismo liviano para identificar el texto generado por un modelo específico, pero su robustez sigue siendo frágil en post-p

La noticia

arXiv:2610.02817v1 Announce Type: new Abstract: Large Language Model (LLM) watermarking provides a lightweight mechanism for identifying text generated by a specific model, but its robustness remains fragile under post-processing attacks. Deletion attacks are particularly challenging because they shift token positions and break the alignment between observed tokens and their original watermark positions. We propose Reed--Muller Code Watermarking (RMCW), an LLM watermarking method based on Reed--Muller codes. In contrast to global codeword recovery, RMCW searches for surviving local algebraic structure, leveraging the Reed--Solomon consistency induced by affine-line restrictions of Reed--Muller codewords. During generation, RMCW injects a Reed--Muller structure into the sequence via a secret-keyed vocabulary partition. During detection, it maps the