EpochLex entry

Pass@k

pass at kay

An evaluation metric estimating whether at least one successful solution appears among k independently sampled model outputs for the same task, commonly used in code-generation evaluation.