Four LLM loss functions → four flavors of LLM misalignment

T2Covered by 1 source · first reported by AI Alignment Forum / LessWrong (curated) at 14 Aug, 22:31 UTC

It seems to me that, for every loss function that we use to train LLMs, we get a very distinct flavor of LLM misalignment. Here’s the summary table, and then we’ll go through the rows separately. Training stage Loss function Flavor of misalignment[1] Famous examples Pretraining & SFT Imitative…

ai

Sources

• T2Four LLM loss functions → four flavors of LLM misalignmentAI Alignment Forum / LessWrong (curated)14 Aug, 22:31