A new framework shows concept subspaces are not unique, estimator choice affects containment and disentanglement, LEACE works well but generalizes poorly, and HuBERT encodes phone info as contained and disentangled from speaker info while speaker info resists compact containment.
Ravel: Evaluat- ing interpretability methods on disentangling language model representations
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2verdicts
UNVERDICTED 2representative citing papers
InfoRidge reveals a non-monotonic pattern in which predictive mutual information between hidden states and outputs peaks in intermediate layers before declining in final layers.
citing papers explorer
-
A framework for analyzing concept representations in neural models
A new framework shows concept subspaces are not unique, estimator choice affects containment and disentanglement, LEACE works well but generalizes poorly, and HuBERT encodes phone info as contained and disentangled from speaker info while speaker info resists compact containment.
-
The Generalization Ridge: Information Flow in Natural Language Generation
InfoRidge reveals a non-monotonic pattern in which predictive mutual information between hidden states and outputs peaks in intermediate layers before declining in final layers.