News for hackers. Stuff that matters.
Re:DeepThought-8B: A small, capable reasoning model by optimalsolver on Saturday November 30, 2024 @05:11AM (#42280733)
Wow, so close. I wonder what went wrong here?
Re:DeepThought-8B: A small, capable reasoning model by brookst on Saturday November 30, 2024 @11:55AM (#42282552)
Tokenization, still. LLMs don’t see individual letters. It’s like using Google translate to ask a German speaker the same question and concluding that Germans can’t count letters.