phenomenon · first seen 12 Oct 2026
model evaluation under reruns
Assessment of fine-tuned model reliability by considering outcome variance across training seeds rather than mean accuracy.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Mean Accuracy Is the Wrong Estimand for a Release DecisionPapers · 12 Oct 2026 · Jaewon Choi, Jihye Kang, Namhyuk Ahn
Latest sources
- Mean Accuracy Is the Wrong Estimand for a Release DecisionPapers · 12 Oct 2026 · Jaewon Choi, Jihye Kang, Namhyuk Ahn