CoVeR: Coverage-Based Token Pruning for Multi-View 3D Reasoning in VLMs
Open paper
Benchmark
Model
Rank
Results
question-answering-on-sqa3d
CoVeR (Video-3D LLM,16views,10%tokens)
#5
AnswerExactMatch (Question Answering): 55.1
Rank counts only results with a code link.