Duwak: Dual Watermarks in Large Language Models

arXiv:2403.13000v1 Announce Type: cross
Abstract: As large language models (LLM) are increasingly used for text generation tasks, it is critical to audit their usages, govern their applications, and mitigate their potential harms. Existing watermark techniques are shown effective in embedding single human-imperceptible and machine-detectable patterns without significantly affecting generated text quality and semantics. However, the efficiency in detecting watermarks, i.e., the minimum number of tokens required to assert detection with significance and robustness against post-editing, is still debatable. In this …

applications arxiv audit critical cs.ai cs.cl cs.cr cs.lg generated govern human language language models large llm machine patterns quality single techniques text watermarks