AI evals are becoming the new compute bottleneck
• 29
AI & ML interests
We’re building a research coalition on evaluating evaluations (EvalEval)! Hosted by Hugging Face, University of Edinburgh, and EleutherAI.
Recent Activity
View all activity
Articles
evaleval/card_backend
Preview • Updated • 21.8k • 1
evaleval/entity-registry-data
Viewer • Updated • 229k • 326
evaleval/EEE_datastore
Updated • 42.6k • 27
evaleval/auto-benchmarkcards
Preview • Updated • 702 • 3
evaleval/alphaxiv_datastore
Updated • 20 • 1
evaleval/social_impact_eval_annotations
Viewer • Updated • 4.24k • 91 • 4
evaleval/old_eee
Preview • Updated • 25