Very good model
I still can’t find a model that would surpass yours in intelligence. She is good at understanding instructions, reasoning adequately, and even playing rp quite well. I've been testing different models with the same riddles for quite some time now. As a rule, even models from Llama3 70B cannot cope with all their tasks. This model, although not the first time and depending on the settings, but gave the correct answers to both riddles ("Today I have 32 apples. Yesterday I ate 28. How many apples do I have?" and "Problem: The book costs one dollar plus half the price of the book. How many dollars does the book cost?".) Other models usually answer only one of them correctly. I also tried to ask other riddles, I think that she passed 50% of my riddles successfully, and if model made a mistake, she only needed a little push in the right direction so that she would find her mistake and answer correctly. I also experimented with the task of finding and writing out objects with which the characters interacted in rp, listing their position in space and state, since sometimes models simply forget where they are and what surrounds them in the scene. I won’t say that it always did everything consistently and correctly, but perhaps I had a complex and overloaded prompt, however, in some cases the model coped with this task simply excellently.
I wonder why this model is not popular, because in my opinion it is really very good. I still use it periodically for various tasks and treat it as a standard, comparable to other models. Thank you for creating such a masterpiece!
No problem dawg