winglian

/

llama-3-8b-256k-PoSE

like 41

I tried long context, short, format no format and not getting coherent results from gguf files from this model.

Not sure if it is due to llama.cpp or the model itself