Current status: Unexpectedly good model.
However it's worth remembering https://huggingface.co./NousResearch/gpt4-x-vicuna-13b
Which have unique way to remember the chat conversation while introducing new facts to the chat.
If the behaviour could somehow end up in any of Mythologic-mini model it would probably make it better and interesting to play around.

Compared to Hermes-1 LLAMA-1 13B model MythoLogic-Mini is not as good at raw detailed story generation, although it's surprisingly close to an average story generator.
It feels like it's going around the story, rather than telling more straightforward story with raw details and completing the story points as asked. But that could as well be part of being a LLAMA-2 model and Hermes-2 not being as good in that as well.

Advanced prompt