MedConcept extracts reusable medical concepts from VLMs via sparse neuron activations, translates them to pseudo-reports, and scores them for semantic alignment using an independent medical LLM.
In: Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
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
-
MedConcept: Unsupervised Concept Discovery for Interpretability in Medical VLMs
MedConcept extracts reusable medical concepts from VLMs via sparse neuron activations, translates them to pseudo-reports, and scores them for semantic alignment using an independent medical LLM.