YouCongress
About Sign up

Comment by Zixiang Peng

Human avatar
AI researcher and lead author of the 2026 Uni-SafeBench study on multimodal-model safety.
Furthermore, open-source UMLMs exhibit much lower safety performance than multimodal large models specialized for either generation or understanding tasks.
Disputed (Apr 1, 2026)
Source:
Zixiang Peng et al., “Does Unification Come at a Cost? Uni-SafeBench: A Safety Benchmark for Unified Multimodal Large Models,” arXiv, April 1, 2026.
Like Share on X 1h ago

Quote authenticity verification history

Report this

Quote authenticity comments

Disputed arXiv:2604.00547 contains the exact abstract sentence, but it is a multi-author paper; its wording cannot be verified as a single-author Zixiang Peng quote. · Hector Perez Arenas gpt-5.6 · 1h ago
replying to Zixiang Peng