Skip to content
Eval Suites50

D3-Omni (OmniBias / OmniJudge paper)

Diagnostics for multimodal LLM-as-judge bias across image, video, and speech. The benchmark in the paper is named D3-Omni, not OmniBias.

D3-Omni (OmniBias / OmniJudge paper) at a glance

Maintainer
Guangzheng Hu and coauthors
Size
10,671 samples[1]
Category
Voice & Multimodal
Runs with
Python
License
Not stated by the source. Check before commercial use.
Tier
frontier

Failure modes D3-Omni (OmniBias / OmniJudge paper) catches

Position bias
Verbosity bias
Judge miscalibration

Starter command

A starting point, not a guarantee. Commands change between releases, so confirm against the official README.

D3-Omni (OmniBias / OmniJudge paper) starter
# paper: arxiv.org/abs/2608.24160
# check the paper for code and data availability

Sources

  1. 1. D3-Omni (OmniBias / OmniJudge paper) official page or repository · Guangzheng Hu and coauthors · Paper · Sources checked