Bileve: Securing Text Provenance in Large Language Models Against Spoofing with Bi-level Signature
–Neural Information Processing Systems
Text watermarks for large language models (LLMs) have been commonly used to identify the origins of machine-generated content, which is promising for assessing liability when combating deepfake or harmful content.
Neural Information Processing Systems
Nov-19-2025, 07:51:26 GMT
- Country:
- North America > United States > Virginia (0.04)
- Genre:
- Research Report > Experimental Study (1.00)
- Industry:
- Information Technology > Security & Privacy (1.00)
- Law (1.00)
- Technology: