Pretty good stuff!
Tested this using the Q8_0_L version.
Disclosure! - I had a larger, more detailed review I'd posted on discord, but I can't find it lmao, so I'm sort of going off memory here.
It does a pretty great job at keeping in character and 'reading the room' in certain situations, and keeping the pace of scenes going for the most part. First impressions were pretty darn good with this one.
It's kind of hit or miss with multiple characters, but that might've been a settings/temperature issue on my side.
It likes to really 'get right to it' when it comes to action and erotic stuff, so much so that sometimes it might break character. It's like the model gets so horny that it forgets to read the character card sometimes lmao.
I haven't tested it with any sort of "longer term" story thing, or stuff with more than 4k context, so i'm not sure how well it does with that.
It really reminded me of Stheno, sorta.
Might need a little more elbow grease here and there to work out some kinks which the character oddness, but it's a solid start.
Awesome to hear, i'm glad it's working well! If you happen to have extra time/bandwidth, it'd be cool if you could do some big context generation on the Q8_0 (non _L) version to compare
So have a convo with Q8_0_L, then swap to Q8_0 and regenerate its last reply and see if you notice any difference in quality