News for hackers. Stuff that matters.

A comment on Kimi Linear: An Expressive, Efficient Attention Architecture. Parent comment

Re:Kimi Linear: An Expressive, Efficient Attention Architecture by idiotsecant on Tuesday July 28, 2026 @12:39PM (#49086504)
wait, is there evidence of this? I've not observed it. It sounds like the kind of thing that I want to be true because it would be hilarious but that makes me suspicious.
[ Parent | on HN ]
Re:Kimi Linear: An Expressive, Efficient Attention Architecture by petu on Tuesday July 28, 2026 @01:23PM (#49087112)
LLMs can't reliably answer what they are (w/o getting that info from system prompt/tool call/etc), so yea.

https://xcancel.com/teortaxesTex/status/2026130112685416881

I think I've seen same happening with some European languages as well.

[ Parent | on HN ]
Re:Kimi Linear: An Expressive, Efficient Attention Architecture by snovv_crash on Tuesday July 28, 2026 @01:47PM (#49087444)
Ask it in Chinese
[ Parent | on HN ]