Agent-SafetyBench: Evaluating the Safety of LLM Agents
This addresses the lack of benchmarks for assessing safety risks in LLM agents, which is crucial for developers and researchers deploying agents in interactive environments, though it is incremental as it focuses on evaluation rather than solving the safety issues directly.
The authors tackled the problem of evaluating the safety of large language model (LLM) agents by introducing Agent-SafetyBench, a comprehensive benchmark with 349 interaction environments and 2,000 test cases, and found that none of the 16 popular LLM agents achieved a safety score above 60%.
As large language models (LLMs) are increasingly deployed as agents, their integration into interactive environments and tool use introduce new safety challenges beyond those associated with the models themselves. However, the absence of comprehensive benchmarks for evaluating agent safety presents a significant barrier to effective assessment and further improvement. In this paper, we introduce Agent-SafetyBench, a comprehensive benchmark designed to evaluate the safety of LLM agents. Agent-SafetyBench encompasses 349 interaction environments and 2,000 test cases, evaluating 8 categories of safety risks and covering 10 common failure modes frequently encountered in unsafe interactions. Our evaluation of 16 popular LLM agents reveals a concerning result: none of the agents achieves a safety score above 60%. This highlights significant safety challenges in LLM agents and underscores the considerable need for improvement. Through failure mode and helpfulness analysis, we summarize two fundamental safety defects in current LLM agents: lack of robustness and lack of risk awareness. Furthermore, our findings suggest that reliance on defense prompts alone may be insufficient to address these safety issues, emphasizing the need for more advanced and robust strategies. To drive progress in this area, Agent-SafetyBench has been released at https://github.com/thu-coai/Agent-SafetyBench/ to facilitate further research in agent safety evaluation and improvement.