Pretraining and alignment induce asymmetric geometric traces in transformer weights because alignment updates concentrate in read pathways due to activation covariance while write pathways inherit less structure from alignment losses.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
fields
cs.LG 2years
2026 2verdicts
UNVERDICTED 2roles
method 1polarities
use method 1representative citing papers
Instruction token embeddings encode visual information that can be leveraged to detect object hallucinations in MLLMs via a new combined score outperforming prior detectors.
citing papers explorer
-
Where Pretraining writes and Alignment reads: the asymmetry of Transformer weight space
Pretraining and alignment induce asymmetric geometric traces in transformer weights because alignment updates concentrate in read pathways due to activation covariance while write pathways inherit less structure from alignment losses.
-
Instruction Lens Score: Your Instruction Contributes a Powerful Object Hallucination Detector for Multimodal Large Language Models
Instruction token embeddings encode visual information that can be leveraged to detect object hallucinations in MLLMs via a new combined score outperforming prior detectors.