Big improvements, but also new issues.
This model is fantastic; it engages with the environment and different aspects of character cards more than any other model I've used, and has far less cliches than in Stheno-v3.2.
However, Lunaris-v1 has its own issues: the response length can sometimes be very long, causing it to go on a rant; it tends to not take direct action, saying that it will do something but never actually doing it; and its performance outside of roleplay took a hit.
I did a simple merge, replacing Stheno-v3.2's weights with Lunaris-v1's but leaving everything else the same, and it seems to have fixed most of the issues while retaining the improvements. You can find it here: https://huggingface.co/HiroseKoichi/L3-8B-Lunar-Stheno
I've tried dozens of other models, but yours have consistently been at the top of my list. I hope you continue making more amazing models in the future!
It is impressive how creative Lunaris v1 is. I do like it more than Stheno 3.2. My issue with Lunaris is that it messed up the context of the roleplay sometimes, even when I tried lowering the temperature to what is recommended (from 1.4 to 0.9).