News for hackers. Stuff that matters.

A comment on LLMs trained on “A is B” fail to learn “B is A”. Parent comment

Re:LLMs trained on “A is B” fail to learn “B is A” by xrisk on Saturday September 23, 2023 @08:22AM (#37622642)
Definitely, but treating fen positions as tokens isn’t going to work, you need something smaller, something you can decompose into. What that is I don’t know.
[ Parent | on HN ]
Re:LLMs trained on “A is B” fail to learn “B is A” by lewhoo on Saturday September 23, 2023 @08:30AM (#37622693)
Well, every character of the string can be a token. From a transformer's perspective it shouldn't matter if it is Chinese, English or "chess" language.
[ Parent | on HN ]