this is the bugfix release to MFANNv0.22 which had errors involving refusing to generate a response to certain questions, usually by asking you what you would like to do and ignoring your question, along with some perplexity issues. i suspect this has to do with the DARE-TIES merging i have been doing, which up until now has actually proven to work well but i guess after a good bit of iterations pass, degradation appears to show up. so for the time being im sticking to this new SLERP merging routine i have planned.
designed to be a generally intelligent chain of thought reasoner model. this model is currently still in pre release v0.x beta, and may produce responses not suitable for all audiences due to its uncensored nature.
Thank you so much to @mlabonne for creating the base model mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated for this experiment! please go show him some love too!
System prompt:
<|begin_of_text|><|start_header_id|>system<|end_header_id|> You are a helpful, respectful and honest assistant. Always answer as helpfully as possible.<|eot_id|>