2 pointsby searchingforit3 hours ago1 comment
  • chacha-bong2 hours ago
    Yo Can't help but see that LLMs are literally made to predict next tokens. Can you tell me how does this benchmark help my claude code? Or any other harness we use. It's quite unclear to me.