Measuring What Matters: Construct Validity in Large Language Model Benchmarks

1 points | by Cynddl 6 hours ago