rish
@rish
Final Friday launch We are upgrading the Neynar score algo - if you're playing mini apps and not writing casts, the previous model scored you poorly; that is getting fixed - if you're writing LLM slop, the new model will be better at catching that - there are some sneakier methods we discovered people were employing, those now are caught as well Scores will recalculate over the weekend and get turned on next week. It's possible there's some volatility before it settles in. If you notice glaring errors, reach out. We want to hear about them and fix them. Linked blog post has more details with analysis. https://neynar.com/blog/retraining-neynar-user-score-algorithm Have a great long weekend!
5 replies
6 recasts
52 reactions
Ghost π©
@ghostbo4.eth
Does Naynar have a Official Token by any chance ?
1 reply
0 recast
0 reaction
rubinovitz
@rubinovitz
Is there mini app usage data somewhere?
1 reply
0 recast
2 reactions
vishal
@vmathur
Whatβs a good score threshold for gating βgoodβ/ high quality users
1 reply
0 recast
0 reaction
Koolkheart
@koolkheart.eth
Love the transparency here. Algo tweaks are tough to get right, but openness like this helps a lot. Curious how are you weighing creativity vs. quantity in the new model?
0 reply
0 recast
0 reaction
Jonathan πποΈπ«
@cellar
Nice!!
0 reply
0 recast
0 reaction