Trending storyDeveloping
Hugging Face policy chief: AI evaluation is hard, and today's methods don't compare or last
1 report1 sourceupdated 1 hour ago
What happened
AI digest
Speaking at TechCrunch Disrupt on October 20, Hugging Face chief policy officer Irene Solaiman said building embedded AI evaluators is not easy. Reproducibility of results depends on open weights, and evaluation is still an evolving science: methods are hard to compare with each other and go stale quickly.
Written by AI from the coverage · updated 53 minutes ago
Coverage
Follow the reports to see the story from different sides.
Oct 20
- TechCrunch · AIEmbedded AI evaluators won’t be an easy solution, Hugging Face’s policy chief says
Hugging Face 首席政策官 Irene Solaiman 在 TechCrunch Disrupt 上表示，嵌入式 AI 评估器并非易事，评估结果的可复现性依赖开放权重，而当前评估方法仍是一门演进中的科学，彼此之间难以有效比较且很快失效。
Heat over time
Not enough continuous observations to draw a trend yet.