2 papers
cs.CL2026
Navigating the Concept Space of Language Models
Wilson E. MarcÃlio-Jr, Danilo M. Eler
Sparse autoencoders (SAEs) trained on large language model activations output thousands of features that enable mapping to human-interpretable concepts. The current practice for an…
cs.LG2025
HUMAP: Hierarchical Uniform Manifold Approximation and Projection
Wilson E. MarcÃlio-Jr, Danilo M. Eler, Fernando V. Paulovich +1
Dimensionality reduction (DR) techniques help analysts to understand patterns in high-dimensional spaces. These techniques, often represented by scatter plots, are employed in dive…