Yeah dolphin yi 34b is better than capypara yi 34b in rp from my biased test too. It’s shame I can’t run goliath on my pc to really suckle that unlimited pseudo GPT4 like experience. But I’m actually rather content with current yi 34b dolphin thanks for insane context size support while still better than any 7b and 13b models.
While I’m only beginning to use dolphin2-yi-34b, it’s giving me good results much consistent and creative than any of 7b or 13b models I’ve used so far! I’ll update the comment when I find something lacking in the future. Your reviews are really helping people like me who don’t have beefy pc or dedication to test tens of different models thoroughly, thank you.
The merging is complete