PerceptionRubrics: Calibrating Multimodal Evaluation to Human Perception
PerceptionRubrics presents a rubric-based evaluation framework that identifies gaps between benchmark scores and real-world performance through atomic auditing and gated scoring me…
Hugging Face · Daily Papers
·Yana Wei, Hongbo Peng
·
·▲ 24 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Yana Wei, Hongbo Peng, Yanlin Lai, Liang Zhao, Kangheng Lin, En Yu
- 24 upvotes da comunidade
- Temas: rubric-based evaluation, atomic auditing, golden captions, Circular Peer-Review consensus, dual-stream system, Must-Right
Resumo
Resumo original (em inglês), extraído do paper:
PerceptionRubrics presents a rubric-based evaluation framework that identifies gaps between benchmark scores and real-world performance through atomic auditing and gated scoring mechanisms.Onde ler
// relacionados
Leia também
Editorial
O gargalo que ninguém via: por que modelos multimodais "colam" durante o treino
Blog
Steal the Patch Size: Adversarially Manipulate Vision-Language Models
Blog
What's Hidden Matters: Identifying Planning-Critical Occluded Agents using Vision-Language Models
Blog