metabench - A Sparse Benchmark of Reasoning and Knowledge in Large Language Models.

Alexander Kipnis, Konstantinos Voudouris, Luca M. Schulze Buschoff, Eric Schulz

Browse the full ICLR paper archive.