Hey!
We are looking for people to join us.
What happens:Why us?
There are a couple of benchmarks that are open source
You can use the training dataset for evaluation
But we belive its not enough.
We want to make newer benchmarks, and more personalized.
We want to design a loop:
- we start on a low level
- partners get better and better
- we make a new benchmark with highter optimization goal
- tiny models keep winning
- we keep giving guidance for their capabilities trough our benches
- repeat the cycle
We want as much partnerships as possible
The more people we can gather, more organized everything will be.
Any questions?
You can start a discussion on your server (discord) or the huggingface disscussions.
Join now
https://huggingface.co/spaces/seton-labs/PartnershipsWith ❤️ for AI capability.