Showing 1–9 of 11 · #evals

Nov 09TESTSHow to test whether an AI model really reasons: the contradiction-trap eval7 min ↗Nov 07MODELSHow to evaluate a new AI model in two weeks: a practical review framework6 min ↗Nov 06BENCHMARKSBuild an LLM eval set in an afternoon6 min ↗Nov 05BENCHMARKSBenchmark contamination: how to check whether a model has seen the test6 min ↗Oct 24BENCHMARKSModel graders are useful, if you calibrate them first6 min ↗Oct 22BENCHMARKSLost in the middle: why long-context models still forget, and how to test it6 min ↗Oct 11AGENTSTool-use evals are broken. Here's how to fix them.6 min ↗Oct 01BENCHMARKSYour 2-point improvement might be noise: confidence intervals for LLM evals6 min ↗Aug 05BENCHMARKSWhen a benchmark saturates, what should replace it?6 min ↗