🧪 Test?View on arXiv
The Unwritten Benchmark: A New Challenge for Multimodal Machine Learning in Abstract Perceptual Reasoning
multimodalreasoningcognitivebenchmark
2608.14558
Builder Relevance
1h ago70%
Abstract
This paper introduces The Unwritten Benchmark, a challenge for multimodal models to perform abstract perceptual reasoning through acousto-kinematic word inference.
Reality Card
Core Claim
Multimodal models like GPT-4o and Gemini 2.5-Pro struggle significantly with acousto-kinematic word inference, achieving less than 10% accuracy compared to over 80% for humans.
Method / Result
Human participants achieve over 80% ordered letter accuracy, while leading models fail to surpass 10%.
Limitations
The models exhibit a paradoxical fusion effect where combining modalities degrades performance, indicating a breakdown in synthesizing perceptual cues.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.