🔎 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

    
def f(s):
    return '{}{}{}'.format(s[3:], s[2], s[5:8])
assert f('jbucwc') == 'cwcuc'
CRUXEval-I 
assert f(??) == 'cwcuc'
 

            		
          

          

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

              

              

            		
          

        

      


      
      

      

        
        
        
        

          414 / 800
        

      


      
      

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