View all comments ➔Show context ➔
  • Fine’ it’s a pLM not an LLM.

    It still uses neural networks and backprop and attention heads. It’s not like it’s radically different computing paradigm FFS.