STRUDEL establishes a scalable foundation for assessing and advancing VLMs torward deeper and more systematic understanding of structured visual information across domains, and reveals that models excel at association tasks, yet struggle with quantification and identification, where precise structural understanding is...
Daniel Steinigen, Lucie Flek, Sebastian Houben· International Conference on...· 0 citations
This paper introduces Explanatory Narratives for CIS, which combine the organizing benefits of storytelling to enable explanation, and proposes ARIC, a cognitively grounded framework for evaluating explanatory narratives across four comprehension stages: Attention, Representation, Integration, and Consolidation.
Vahid Sadiri Javadi, Sadia Naseer, Ali Ather et al.· International Conference on...· 0 citations
It is shown that rewarding narrow, overtly misaligned behavior produces substantially higher general-domain misalignment than sample-matched SFT, and that EM from RL can be induced by reward signals that could plausibly arise naturally, such as unpopular aesthetic preferences or poor rhetorical appeals.
Magnus Jørgenvåg, David Kaczér, Lasse Ruttert et al.· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.