VIBE benchmark evaluates visual instruction following in image editing models across deictic, morphological, and causal levels, finding proprietary models lead but all degrade on harder tasks.
- Do NOT list differences caused by: • minor blur or softness, • small texture or color shifts, • pixel-level noise, • slight position or alignment offsets
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
How Well Do Models Follow Visual Instructions? VIBE: A Systematic Benchmark for Visual Instruction-Driven Image Editing
VIBE benchmark evaluates visual instruction following in image editing models across deictic, morphological, and causal levels, finding proprietary models lead but all degrade on harder tasks.