Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Explanation Generation on VQA-X

43.9Preference Rate (Ours)

OFA-X

Updated 1mo ago

Evaluation Results

Method	Links
OFA-X 2022.12		43.9	25	31.1

SOTA Paper

OFA-X

Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations

Dataset

VQA

Follow for update

@wizwand_team Discord

Related Benchmarks

Visual Question Answering on VQA-X e-ViL (test)Visual Entailment on e-SNLI-VE e-ViL (test)Visual Commonsense Reasoning on VCR e-ViL (test)Explanation Generation on e-SNLI-VE Explanation Generation on VCR

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs