This is a ties merge between
Nous-Hermes-2-Mistral-7B-DPO,
Noromaid-7B-0.4-DPO and
multi_verse_model. Config included. Seems extremely capable in RP (and still quite okay in other tasks) for 7B model. The main idea was to take astonishingly powerful multi_verse_model and make it a bit more obedient. In the process I found out that preferred prompt format has shifted from alpaca to chatml.
