Task-specific LoRA adapters in continual learning exhibit significant low-rank subspace overlap, enabling LiteLoRA's learned gating to reduce active adapters by 20-70% while matching or exceeding prior performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
When One Adapter Speaks for Many: Discovering Low-Rank Redundancy in Continual Fine-Tuning
Task-specific LoRA adapters in continual learning exhibit significant low-rank subspace overlap, enabling LiteLoRA's learned gating to reduce active adapters by 20-70% while matching or exceeding prior performance.