Concept
best@K
best@K measures the quality you get when you generate K candidate outputs for a single problem and keep the best one according to some selection rule. It is the test-time scaling cousin of pass@K, with one important difference.
The rest of “best@K” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→