Lost or Hidden? A Concept-Level Forgetting in Supervised Continual Learning

In three linesStudy of forgetting in continual learning using Sparse Autoencoders (SAEs). Authors propose a diagnostic framework to analyze how task-specific information evolves at concept-level granularity. Finding: much apparent forgetting stems from representational accessibility loss rather than complete information erasure.

Summary generated by Claude — human-verified