Introduces Abstraction Gap metric and CAGE benchmark showing seven of eight VLMs have large gaps between text plausibility and chain-based causal reasoning, with one model succeeding.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Abstraction Gap in Vision-Language Causal Reasoning
Introduces Abstraction Gap metric and CAGE benchmark showing seven of eight VLMs have large gaps between text plausibility and chain-based causal reasoning, with one model succeeding.