4 ms·
It seems straightforward that finetuning an LLM just for usefulness will reduce alignment (and the other way around).
by shikon7 3y ago
It seems straightforward that finetuning an LLM just for usefulness will reduce alignment (and the other way around).
- pacchu78 3y agoYes, that was the initial hypothesis, which has been proven and shown conclusively to be true. If this was straightforward, then we should be more careful and do more tests when using fine-tuned models.