12 citations · 16 across the 15 of their papers we have counts for
3 papers · 1 filter
ResiDual Transformer Alignment with Spectral Decomposition
Lorenzo Basile, Valentino Maiorca, Luca Bortolussi +2
When examined through the lens of their residual streams, a puzzling property emerges in transformer networks: residual contributions (e.g., attention heads) sometimes specialize i…
Intrinsic Dimension Correlation: uncovering nonlinear connections in multimodal representations
Lorenzo Basile, Santiago Acevedo, Luca Bortolussi +2
To gain insight into the mechanisms behind machine learning methods, it is crucial to establish connections among the features describing data points. However, these correlations o…
When Can You Trust Your Explanations? A Robustness Analysis on Feature Importances
Ilaria Vascotto, Alex Rodriguez, Alessandro Bonaita +1
Recent legislative regulations have underlined the need for accountable and transparent artificial intelligence systems and have contributed to a growing interest in the Explainabl…