A zero-shot MLLM pipeline that objectifies safety rules, decomposes them into preconditions, and uses debiased token probabilities plus cascaded reasoning achieves 94.8% accuracy on a synthetic image safety benchmark.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MLLM-as-a-Judge for Image Safety without Human Labeling
A zero-shot MLLM pipeline that objectifies safety rules, decomposes them into preconditions, and uses debiased token probabilities plus cascaded reasoning achieves 94.8% accuracy on a synthetic image safety benchmark.