Dedicated to the study of techniques for embedding digital watermarks in the outputs of language models to ensure authenticity and traceability of generated content.