How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures
Existing vision-language model (VLM) benchmarks emphasize perception and reasoning accuracy (how well VLMs describe and reason about what they see in an image), with limited attention to behavioral reliability under uncertainty (how they behave when visual evidence is missing or misleading). We introduce SciFigBench, a diagnostic VLM benchmark for scientific figure understanding that jointly evaluates perception, reasoning, and behavioral reliability under uncertainty. It contains 250 figures with high-quality human annotations across three evaluation aspects, totaling 600+ hours of annotation
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- LinkedLinked via arxiv author · 85%Paul Osemudiame Oamen →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Owusu-Banahene Osei →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Ananya Mukherjee →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Christian Greisinger →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Steffen Eger →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Pius Onobhayedo →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”
- LinkedLinked via arxiv author · 85%Wei Zhao →
“How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures”