Sparse autoencoders on ViT class tokens reveal stable Class Activation Profiles for in-distribution data, enabling OOD detection via divergence from core energy profiles.
Ex- ploring the limits of out-of-distribution detection.Advances in neural information processing systems, 34:7068–7081
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
An adapted Cellpose-SAM pipeline achieves 1.50% MAPE on ASTM grain size number G using only two training images while maintaining topological separation better than U-Net, MatSAM, or Qwen2.5-VL-7B.
citing papers explorer
-
Sparsity as a Key: Unlocking New Insights from Latent Structures for Out-of-Distribution Detection
Sparse autoencoders on ViT class tokens reveal stable Class Activation Profiles for in-distribution data, enabling OOD detection via divergence from core energy profiles.
-
Bridging Foundation Models and ASTM Metallurgical Standards for Automated Grain Size Estimation from Microscopy Images
An adapted Cellpose-SAM pipeline achieves 1.50% MAPE on ASTM grain size number G using only two training images while maintaining topological separation better than U-Net, MatSAM, or Qwen2.5-VL-7B.