Hacker News new | ask | show | jobs
by coder543 911 days ago
I wish that Arena included a few more "interesting" models like the new Phi-2 model and the current tinyllama model, which are trying to push the limits on small models. Solar-10.7B is another interesting model that seems to be missing, but I just learned about it yesterday, and it seems to have come out a week ago, so maybe it's too new. Solar supposedly outperforms Mixtral-8x7B with a fraction of the total parameters, although Solar seems optimized for single-turn conversation, so maybe it falls apart over multiple messages (I'm not sure).
2 comments

Solar-10.7B is present in the battle arena but there are probably not enough votes for the ranking.
> like the new Phi-2 model

Phi-2 isn't fine tuned for instruction following yet.