The Trust Paradox: How CS Researchers Engage LLM Leaderboards
Signal
72
Hype
15
In three linesQualitative study of 8 AI researchers reveals a paradox: they distrust LLM leaderboards yet use them as decision aids. Peer networks dominate model selection. NLP researchers face SOTA pressure absent in HCI/Systems. Universal demand: cost transparency.Read source
Your take?
Summary generated by Claude — human-verified