Avijit Ghosh publishes a new papers update.

I do hope this changes! For eg a lot of tiny orgs and academic groups are going high quality experiments that don’t see life beyond a paper/dataset because it takes a massive amount of effort/capital to have a sustained eval over several model generations.