🔎 CRUXEval Sample Explorer 🔎

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


   
def f(text):
    new_text = []
    for character in text:
        if character.isupper():
            new_text.insert(len(new_text) // 2, character)
    if len(new_text) == 0:
        new_text = ['-']
    return ''.join(new_text)
assert f('String matching is a big part of RexEx library.') == 'RES'
CRUXEval-I
assert f(??) == 'RES'

CRUXEval-O
assert f('String matching is a big part of RexEx library.') == ??