IxinDow

joined 10 months ago
sorted by: new top controversial old
Training LLMs to follow procedure for Math gives an accuracy of 98.5% in c/localllama@poweruser.forum
[โ€“] IxinDow@alien.top 1 points 10 months ago

Can you try one of Yi-34B's variants?

Introducing Tess: Tess-M with 200K Context Length in c/localllama@poweruser.forum
[โ€“] IxinDow@alien.top 1 points 10 months ago

How many tokens in your substack example?
Do you have examples of using model for fiction with length 16K-40K tokens?

1
He must be very enlightened in using LLM (alien.top)