@hjalexnam
CS PhD student at Stanford University
AI
The researchers say it generates tasks from existing benchmarks and uses them to diagnose failures in AI tool use.
AI
