List your company

Metric Sensitivity Analysis on Quilt-1M Logic Error

91ScoreBERTScore
9.8830.945273.06Mar 17, 2026

Evaluation results

MethodLinks
BERTScoreFocus=Semantic2026.03911.1
PathGLS (Sg)Focus=Visual-Text2026.03735.2
PathGLS (Sl)Focus=Consistency2026.036726.4
RadGraphFocus=Entity2026.032519.4
BLEU-4Focus=Lexical2026.031318.8