CredID: Credible Multi-Bit Watermark for Large Language Models Identification

2024-12-04

arXiv.org

AI Summary

Large language models (LLMs) pose privacy and security risks due to the absence of identity attribution mechanisms, and existing text watermarking methods struggle to simultaneously achieve high output quality, sufficient capacity, and robustness. Method: We propose the first trusted multi-party watermarking framework, introducing a trusted third party (TTP) to coordinate multiple model vendors for watermark embedding and joint verification—without exposing user prompts. Our approach features a multi-bit adaptive watermark encoding scheme and a key-derivation seed mechanism, preserving vendor data privacy while enhancing capacity (supporting >16 concurrent vendors), robustness (resilient against cross-model obfuscation), and trustworthiness. A lightweight algorithm ensures minimal impact on text quality (perplexity increase <0.8%) and achieves >99.2% watermark detection accuracy. Contribution/Results: We open-source the implementation to facilitate reproducible research.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Large Vision Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Large Language Models (LLMs) are widely used in complex natural language processing tasks but raise privacy and security concerns due to the lack of identity recognition. This paper proposes a multi-party credible watermarking framework (CredID) involving a trusted third party (TTP) and multiple LLM vendors to address these issues. In the watermark embedding stage, vendors request a seed from the TTP to generate watermarked text without sending the user's prompt. In the extraction stage, the TTP coordinates each vendor to extract and verify the watermark from the text. This provides a credible watermarking scheme while preserving vendor privacy. Furthermore, current watermarking algorithms struggle with text quality, information capacity, and robustness, making it challenging to meet the diverse identification needs of LLMs. Thus, we propose a novel multi-bit watermarking algorithm and an open-source toolkit to facilitate research. Experiments show our CredID enhances watermark credibility and efficiency without compromising text quality. Additionally, we successfully utilized this framework to achieve highly accurate identification among multiple LLM vendors.

Problem

• Addressing LLM identity recognition for privacy and security
• Developing multi-bit watermarking with high capacity and robustness
• Enabling credible multi-vendor identification without privacy compromise

Innovation

• Multi-party watermarking framework with TTP
• Novel multi-bit watermarking algorithm
• Open-source toolkit for research facilitation