WISE

Dataset Information
Introduced
2025
License
MIT
Homepage

Overview

WISE, the first benchmark specifically designed for World Knowledge-Informed Semantic Evaluation. WISE moves beyond simple word-pixel mapping by challenging models with 1000 meticulously crafted prompts across 25 sub-domains in cultural common sense, spatio-temporal understanding, and natural science.

Variants: WISE

Associated Benchmarks

This dataset is used in 1 benchmark:

Recent Benchmark Submissions

Task Model Paper Date
Image Generation UniWorld-V1 UniWorld-V1: High-Resolution Semantic Encoders for … 2025-06-03
Image Generation Bagel (w/ cot) Emerging Properties in Unified Multimodal … 2025-05-20
Image Generation Bagel Emerging Properties in Unified Multimodal … 2025-05-20
Image Generation MindOmni (w/o cot) MindOmni: Unleashing Reasoning Generation in … 2025-05-19
Image Generation MindOmni (w/ cot) MindOmni: Unleashing Reasoning Generation in … 2025-05-19
Image Generation Janus Janus-Pro: Unified Multimodal Understanding and … 2025-01-29
Image Generation Janus-pro Janus-Pro: Unified Multimodal Understanding and … 2025-01-29
Image Generation Emu3-gen Emu3: Next-Token Prediction is All … 2024-09-27
Image Generation Show-o Show-o: One Single Transformer to … 2024-08-22
Image Generation stable-diffusion-3.5-large Scaling Rectified Flow Transformers for … 2024-03-05
Image Generation Playground-v2.5-1024px-aesthetic Playground v2.5: Three Insights towards … 2024-02-27
Image Generation PixArt-XL-2-1024-MS PixArt-$α$: Fast Training of Diffusion … 2023-09-30
Image Generation stable-diffusion-xl-base-0.9 SDXL: Improving Latent Diffusion Models … 2023-07-04

Research Papers

Recent papers with results on this dataset: