🔎 CRUXEval Sample Explorer 🔎

CRUXEval is a benchmark complementary to HumanEval and MBPP measuring code reasoning, understanding, and execution capabilities!


   
def f(orig):
    copy = orig
    copy.append(100)
    orig.pop()
    return copy
assert f([1, 2, 3]) == [1, 2, 3]
CRUXEval-I
assert f(??) == [1, 2, 3]

CRUXEval-O
assert f([1, 2, 3]) == ??