1,595 Downloads Updated 8 months ago
This is an uncensored version of cognitivecomputations/Dolphin3.0-R1-Mistral-24B created with abliteration (see remove-refusals-with-transformers to know more about it).
This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.
bc1qqnkhuchxw0zqjh2ku3lu4hq45hc6gy84uk70ge