HumanEval

HumanEval: A benchmark of one hundred and sixty-four Python programming problems that scores whether generated code passes hidden unit tests.

Browse the full glossary →