Red-Teaming

T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search
While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models (LLMs), such approaches fail …
Hyomin Lee, Sangwoo Park, Yumin Choi, Sohyun An, Seanie Lee, Sung Ju Hwang
Cite Paper Code