OmniAlign-V: Towards Enhanced Alignment of MLLMs with Human Preference
Xiangyu Zhao, Shengyuan Ding, Zicheng Zhang, Haian Huang et al.
43.60/100
💧 Filler
Padding, dubious value
Content 43.6 · Citation bonus +0.0 · no citation data
💡 This paper proposes the 200K multi-modal preference alignment dataset OmniAlign-V and the human-annotated alignment benchmark MM-AlignBench, validates that SFT/DPO can improve MLLMs' human preference
#多模态对齐基建#数据堆料#DPO again#开源交朋友#multi-modal alignment in#dataset stacking#DPO again#open-source friendship
Score breakdown
Novelty3.0 / 10
Rigor5.0 / 10
Significance6.0 / 10
Clarity8.0 / 10
Reproducibility7.0 / 10
This tone hasn't been generated yet — roast it again to create it.