A theory-grounded EMI metric and an upper bound on its distribution-shift gap are proposed for evaluating multimodal LLM robustness.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Understanding Multimodal LLMs Under Distribution Shifts: An Information-Theoretic Approach
A theory-grounded EMI metric and an upper bound on its distribution-shift gap are proposed for evaluating multimodal LLM robustness.