🔎 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(text, x):
    if text.removeprefix(x) == text:
        return f(text[1:], x)
    else:
        return text
assert f("Ibaskdjgblw asdl ", "djgblw") == 'djgblw asdl '

CRUXEval-I assert f(??) == 'djgblw asdl '

CRUXEval-O assert f("Ibaskdjgblw asdl ", "djgblw") == ??

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