Watermarking Language Models through Language Models

Agnibh Dasgupta, Abdullah Tanvir, Xin Zhong

Watermarking the outputs of large language models (LLMs) is critical for provenance tracing, content regulation, and model accountability. Existing approaches often rely on access…