sigmoid.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A social space for people researching, working with, or just interested in AI!

Server stats:

594
active users

Leshem Choshen

New efficient eval results

1. A few examples are enough for Human preference to be clear, automatic metrics also don't need too many
2. Context may change which model is preferred

arxiv.org/abs/2402.18756

The findings on the redundancy in the amount of examples we use (for humans) complement several recent works showing it for benchmarks:

sigmoid.social/@LChoshen/11097

Finding better ways to choose examples for efficiency of choosing between a pair of models
sigmoid.social/@LChoshen/11192

& human based efficiency (like low resource annotation and active learning things)

label-sleuth.org/

I am sure there are also related work about the context dependency, and other related threads I didn't know, please share (sorry for only mentioning my works, know others?)

www.label-sleuth.orgLabel SleuthOpen-source no-code system for text annotation and building of text classifiers