Loading profile…
ORGANIZATION
@valen-team
Open research project publishing the Valen multimodal decision models; the code lives on the GitHub account Liuziyu77 (Ziyu Liu), and no company is named as the maker.
A multimodal decision model from the Valen team: Qwen3.5-4B plus a two-layer MLP-Mixer decision head that returns probability distributions for Choice, Noul and Score questions about a shared state of text, images or video, without generating answer text.
83.1%4.5Bapache-2.0
