🔎 CRUXEval Sample Explorer 🔎

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

Homepage Paper Code HF Dataset Leaderboard

model: temp:

def f(s):
    left, sep, right = s.rpartition('.')
    new = sep.join([right, left])
    _, sep, _ = new.rpartition('.')
    return new.replace(sep, ', ')
assert f('galgu') == ', g, a, l, g, u, '

CRUXEval-I assert f(??) == ', g, a, l, g, u, '

CRUXEval-O assert f('galgu') == ??

The site is inspired by and based on Minerva sample explorer and Llemma. Thanks!