🔎 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(txt, marker):
    a = []
    lines = txt.split('\n')
    for line in lines:
        a.append(line.center(marker))
    return '\n'.join(a)
assert f('#[)[]>[^e>\n 8', -5) == '#[)[]>[^e>\n 8'
CRUXEval-I 
assert f(??) == '#[)[]>[^e>\n 8'
 

            		
          

          

            
              CRUXEval-O
              
assert f('#[)[]>[^e>\n 8', -5) == ??

              

              

            		
          

        

      


      
      

      

        
        
        
        

          475 / 800
        

      


      
      

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