fahadhafeezofficial/agent-failure-bench download history

fahadhafeezofficial/agent-failure-bench is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 426 times (143 in the last 7 days), and 426 times in total. It ranks #37,529 among datasets by monthly downloads.

AgentFailureBench A synthetic, fully-annotated dataset of multi-step AI-agent tool-use trajectories, built to train and evaluate agent reliability: tool-selection correctness, argument validity, constraint-following, error recovery, and final-answer faithfulness. Dataset s

Open fahadhafeezofficial/agent-failure-bench on Hugging Face