Fetching the paper…

Testing and Understanding Erroneous Planning in LLM Agents through Synthesized User Inputs · Around