Visual Question Answering on ViQuAE
63.17F1 ScoreRC Baseline
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| RC BaselineType=RC Baseline, Evaluation Protocol=Few-shot (full-oracle)2026.03 | 63.17 | 57.55 | — | |
| Wiki-R1Backbone Size=7B2026.03 | 55.6 | 50.3 | — | |
| Wiki-R1Backbone Size=3B2026.03 | 53.8 | 48.6 | — | |
| RC BaselineType=RC Baseline, Evaluation Protocol=Few-shot (semi-oracle)2026.03 | 44.1 | 40.32 | — | |
| RC BaselineType=RC Baseline, Evaluation Protocol=Few-shot2026.03 | 25.43 | 22.07 | — | |
| ReflectiVALLM=LLaMA-3.1-8B, Zero-shot=true2024.11 | 23.2 | 38.1 | 52 | |
| ReflectiVAType=MLLM Baseline2026.03 | 23.2 | 38.1 | — | |
| RC BaselineType=RC Baseline, Evaluation Protocol=Zero-shot2026.03 | 20.96 | 18.06 | — | |
| ReflectiVA (w/o KB)LLM=LLaMA-3.1-8B, Zero-shot=true, Knowledge Base=None2024.11 | 16.6 | 27.6 | — | |
| LLaVA-v1.5LLM=Vicuna-7B, Zero-shot=true2024.11 | 15.1 | 26.6 | — | |
| LLaVA-v1.5Type=MLLM Baseline2026.03 | 15.1 | 26.6 | — | |
| LLaVA-v1.5LLM=LLaMA-3.1-8B, Zero-shot=true2024.11 | 15 | 25.6 | — | |
| Wiki-LLaVA (InfoSeek)LLM=LLaMA-3.1-8B, Zero-shot=true2024.11 | 12.7 | 21.8 | — | |
| Wiki-LLaVAType=MLLM Baseline, Training Data=InfoSeek2026.03 | 12.7 | 21.8 | — | |
| Wiki-LLaVA (E-VQA)LLM=LLaMA-3.1-8B, Zero-shot=true2024.11 | 10.5 | 16.7 | — | |
| Wiki-LLaVAType=MLLM Baseline, Training Data=E-VQA2026.03 | 10.5 | 16.7 | — |