dongsheng/ToolMaze download history

dongsheng/ToolMaze is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 889 times (109 in the last 7 days), and 2,881 times in total. It ranks #21,114 among datasets by monthly downloads.

ToolMaze This project is from the paper "When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents ". Evaluation framework for testing LLM agents on C1/C2/C3/C4 tool-use tasks under P0–P4 perturbation modes. The framework runs an agent inside a sandboxed too

Open dongsheng/ToolMaze on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.