OutSafe-Bench supplies the first large-scale four-modality safety dataset and evaluation framework that exposes persistent unsafe outputs in nine leading multimodal LLMs.
Multitrust: A comprehensive bench- mark towards trustworthy multimodal large language mod- els.Advances in Neural Information Processing Systems, 37:49279–49383
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
OutSafe-Bench: A Benchmark for Multimodal Offensive Content Detection in Large Language Models
OutSafe-Bench supplies the first large-scale four-modality safety dataset and evaluation framework that exposes persistent unsafe outputs in nine leading multimodal LLMs.