phu-pmh/irt-leaderboard

Leaderboards are widely used in NLP and push the field forward. While leaderboards are a straightforward ranking of NLP models, this simplicity can mask nuances in evaluation items (examples) and subjects (NLP models). Rather than replace leaderboards, we advocate a re-imagining so that they better highlight if and where progress is made. Building on educational testing, we create a Bayesian leaderboard where latent subject skill and latent item difficulty predict correct responses. Using this model, we analyze the reliability of leaderboards. Afterwards, we show the model can guide what annotate, identify annotation errors, detect overfitting, and identify informative examples.

PythonApache-2.0