[tdullien]

Author here. I am entirely ok with using "goal" in the context of an RL algorithm. If you read my article carefully, you'll find that I object to the use of "goal" in the context of LLMs.

[Timwi]

If you read the literature on AI safety carefully (which uses the word “goal”), you'll find they're not talking about LLMs either.

[tdullien]

I think the Anthropic "omg blackmail" article clearly talks about both LLMs and their "goals".