martin ↑
@martin
ok need to refine the 100 $higher @aethernet giveaway to make it better i don't think it's rewarding the most interesting conversations right now if you have experience with helping an ai "judge" thing and giving them score, would appreciate insights. the normal llm stuff isn't really working here i think
10 replies
1 recast
45 reactions
JB Rubinovitz ⌐◨-◨
@rubinovitz
Yeah we should label what we consider “interesting conversations” and then higher can use that You can also do this based on likes, likes from higher community members, etc Happy to DM on this too
2 replies
0 recast
9 reactions
martin ↑
@martin
yeah i think that part is what's needed, some sort of labeling system that can then be used
3 replies
0 recast
0 reaction
JB Rubinovitz ⌐◨-◨
@rubinovitz
I think you also need to decide on “interesting to who” Does interesting = viral Or Interesting to the existing higher community, WC, etc
0 reply
0 recast
1 reaction
JB Rubinovitz ⌐◨-◨
@rubinovitz
Happy to talk supervised learning and labeling systems Monday if you end up needing a hand. Stuck in traffic right now.
0 reply
0 recast
0 reaction
Les Greys
@les
If you figure this out you got a secret sauce. Working on something similar but different area. All I can offer is text length can be one, cosine similarity score between aethernet and other account, also frequency within windows of time. Creating scores individually, and seeing patterns on the aggregate may be helpful.
0 reply
0 recast
0 reaction