Lost in Perception: Isolating Perceptual and Reasoning Failures in Multimodal Physics and Geometry Reasoning
A five-task diagnostic experiment across physics and geometry benchmarks that isolates failures to perception, reasoning, or both is introduced, and InternS1-mini, despite heavy scientific pretraining and thinking capabilities, falls below the weakest model from experiments on every task.
Raj Jaiswal, Sree Krishna Uppalapati, Dhruvkumar Patel et al.
· 0 citations