We have released a D-LLM-mutated model of Llama-2-7b-chat on an anonymous account on huggingface.
We have also released our training dataset for ΔW on the same page of huggingface.