un poster criminalmente sottovalutato considerando la quantità di contenuti di qualità e/o stimolanti che pubblica
mike64_t
mike64_t18 ott, 12:33
I think the observation that LLMs are "bad tutors" in that they cannot precisely probe understanding is accurate. The fact that "upweighting the entire rollout" is stupid is also true. However its not obvious to me that the remedy for that is LLM-reflection as to "what went well". I think this runs into very similar issues of collapse-risk or misallocation of supervision. Because while we might be sucking supervision through a straw, the only thing thats even worse is sucking tainted supervision through a straw.
non è che Mike sia un poster di nicchia o qualcosa del genere, ma sto solo pensando a quanti account di merda hanno 2-10 volte di più
30,79K