easy5 pts · first answer countsllm-evaluationbenchmarkscode-eval

How HumanEval scores code

How does the HumanEval benchmark decide if a generated solution is correct?

A.A human reads each solution and gives a gradeB.It compares the code text with a reference using BLEUC.It checks that the code compiles, nothing moreD.It runs hidden unit tests against the generated function

Discussion 0

Join free to post in the discussion.

How HumanEval scores code

Discussion 0

Related questions