Re: [ot][spam][log] attempt bits: talk to a language model on list

- the bloke example code failed to import torch, so i added “import torch” at the top - the default transformers code does not support the context length scaling feature. i found working code at https://github.com/huggingface/transformers/issues/25957#issuecomment-170621... . the working code, sadly, is paired with a comment that the output is poor quality. maybe i’ll use some of this energy to glance at the paper.