Return to Article Details
The Benchmark Trap: Structures of Power and Injustice in AI Evaluations
Download
Download PDF