Optimizing watermarks for large language models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Wouters, Bram
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914624331317248
author Wouters, Bram
author_facet Wouters, Bram
contents With the rise of large language models (LLMs) and concerns about potential misuse, watermarks for generative LLMs have recently attracted much attention. An important aspect of such watermarks is the trade-off between their identifiability and their impact on the quality of the generated text. This paper introduces a systematic approach to this trade-off in terms of a multi-objective optimization problem. For a large class of robust, efficient watermarks, the associated Pareto optimal solutions are identified and shown to outperform the currently default watermark.
format Preprint
id arxiv_https___arxiv_org_abs_2312_17295
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Optimizing watermarks for large language models
Wouters, Bram
Cryptography and Security
Artificial Intelligence
Computation and Language
With the rise of large language models (LLMs) and concerns about potential misuse, watermarks for generative LLMs have recently attracted much attention. An important aspect of such watermarks is the trade-off between their identifiability and their impact on the quality of the generated text. This paper introduces a systematic approach to this trade-off in terms of a multi-objective optimization problem. For a large class of robust, efficient watermarks, the associated Pareto optimal solutions are identified and shown to outperform the currently default watermark.
title Optimizing watermarks for large language models
topic Cryptography and Security
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2312.17295