Three bricks to consolidate watermarks for large language models - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Three bricks to consolidate watermarks for large language models

Résumé

Discerning between generated and natural texts is increasingly challenging. In this context, watermarking emerges as a promising technique for ascribing text to a specific generative model. It alters the sampling generation process to leave an invis- ible trace in the output, facilitating later detection. This research consolidates watermarks for large language models based on three theoretical and empirical considerations. First, we introduce new statistical tests that offer robust theoretical guarantees which remain valid even at low false-positive rates (less than 10-6). Second, we compare the effectiveness of watermarks using classical benchmarks in the field of natural language processing, gaining insights into their real-world applicability. Third, we develop advanced detection schemes for scenarios where access to the LLM is available, as well as multi-bit watermarking.
Fichier principal
Vignette du fichier
fernandez_wifs2023.pdf (773.99 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04361015 , version 1 (22-12-2023)

Licence

Paternité

Identifiants

  • HAL Id : hal-04361015 , version 1

Citer

Pierre Fernandez, Antoine Chaffin, Karim Tit, Vivien Chappelier, Teddy Furon. Three bricks to consolidate watermarks for large language models. WIFS 2023 - IEEE International Workshop on Information Forensics and Security, Dec 2023, Nuremberg, Germany. pp.1-9. ⟨hal-04361015⟩
54 Consultations
31 Téléchargements

Partager

Gmail Facebook X LinkedIn More