✅ Do use LLMs for

  • Double-checking your grammar and wording
    • Do NOT ask language models to completely re-word your text, it will add language that is clearly non-human.

❌ Do NOT use LLMs for

  • Rating the model responses
    • These prompts are intentionally hard enough to where even the best models will make mistakes.
  • Writing rationales / explanations
    • We’re looking for your explanation!
  • Evaluating how code will run
    • Models are generally horrendous at doing this for anything non-trivial.

It’s usually pretty obvious what is generated by AI and what is generated by you, it’s important to follow this guidance. When in doubt, ask us!