RE: RE: How I Get Frontier Results With Local LLMs
You are viewing a single comment's thread from:

RE: How I Get Frontier Results With Local LLMs

Words
93
Reading
1 min
19d

They will do a bunch, but you have to be instructive and keep it on track. The context issue is why I do things in small batches and actually work with it instead of trying to one shot things. I played with Gemma, but got way better results with Qwen. To get them to actually do the work, you usually have to turn off thinking and just make it do the work you tell it to do. I had to do with with frontier models too to get anything useful out of them.

@thelogicaldude: They will do | Ecency