1I used RL fine-tuning to make an LLM generate ugly and unpythonic FizzBuzz code (opens in new tab)(seantey.github.io)4seanrrr3mo ago1