MANSU achieves the first unlearning method that jointly delivers behavioral forgetting, retain preservation, zero PTQ gap, and structural erasure by circuit-restricted null-space projection plus a per-parameter magnitude floor.
Mass-editing memory in a transformer
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
LLMs form functional subspaces in activation space where in-context learning tasks are solved by vector algebra operations such as addition and subtraction.
citing papers explorer
-
Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution
MANSU achieves the first unlearning method that jointly delivers behavioral forgetting, retain preservation, zero PTQ gap, and structural erasure by circuit-restricted null-space projection plus a per-parameter magnitude floor.
-
Functional Subspace, where language models can use vector algebra to solve problems
LLMs form functional subspaces in activation space where in-context learning tasks are solved by vector algebra operations such as addition and subtraction.