作者位次 #4/8ACL 2026Poster· Natural Language Processing
A gent M ark: Utility-Preserving Behavioral Watermarking for Agents
Kaibo Huang · Jin Tan · Yukun Wei · Wanling Li · Zipei Zhang · Hui Tian · Zhongliang Yang · Linna Zhou
摘要
LLM-based agents are increasingly deployed to autonomously solve complex tasks, raising urgent needs for IP protection and regulatory provenance. While content watermarking effectively attributes LLM-generated outputs, it fails to directly identify the high-level planning behaviors (e.g., tool an...