The paper shows existing multimodal safety benchmarks leak harmful image content into text queries (VSIL), and introduces VLSBench, a 2.2k-pair leakless benchmark on which textual alignment fails and multimodal alignment is needed.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
VLSBench: Unveiling Visual Leakage in Multimodal Safety
The paper shows existing multimodal safety benchmarks leak harmful image content into text queries (VSIL), and introduces VLSBench, a 2.2k-pair leakless benchmark on which textual alignment fails and multimodal alignment is needed.