List your company

Video Question Answering on MSVD-QA zero-shot

73.5AccuracyOracle-RLAIF
50.72456.63762.5568.463Oct 2, 2025

Evaluation results

MethodLinks
Oracle-RLAIFFine-tuning method=RLAIF (GRPOrank), Evaluation Mode=Zero-shot2025.1073.54
Video-LLaVAFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1070.73.9
VideoChat2Fine-tuning method=SFT, Evaluation Mode=Zero-shot2025.10703.9
LLaMA-VIDFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1069.73.7
VLM-RLAIFFine-tuning method=RLAIF (PPO), Evaluation Mode=Zero-shot2025.1068.53.6
BT-AdapterFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1067.53.7
VLM-SFTFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1067.23.6
Video-ChatGPTFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1064.93.3
Oracle-DPOFine-tuning method=RLAIF (DPO), Evaluation Mode=Zero-shot2025.1064.73.4
VideoLLaMAFine-tuning method=SFT, Evaluation Mode=Zero-shot2025.1051.62.5