Nivya Talokar

Helpful to a Fault: Measuring Illicit Assistance in Multi-Turn, Multilingual LLM Agents

LLM-based agents increasingly execute real-world workflows via tools and memory. Granting LLMs such powers enables ill-intended adversaries to likewise use these agents to carry out complex misuse scenarios. Existing agent-misuse benchmarks largely test single-prompt instructions, leaving a gap in m…