paper
When to Write and When to Suppress: Route-Specialized Dual Adapters for Memory-Assisted Knowledge Editing
paperactiveprovisional
when-to-write-and-when-to-suppress-route-specialized-dual-adapters-for-memory-assisted-knowledge-editing-3889da67·1 events·first seen 2d agoAliases: When to Write and When to Suppress: Route-Specialized Dual Adapters for Memory-Assisted Knowledge Editing
Co-occurring entities
More like this (12)
knowledge editingKnowledge Editing in Masked Diffusion Language ModelsTrace Only What You Need: Structure-Aware On-Demand Hypergraph Memory for Long-Document Question AnsweringRecalling Too Well: Sycophancy Evaluation and Mitigation in Memory-Augmented Modelslanguage-aware adapter headsCounterfactual EditingSemantic Edit RegularizationAttention Amnesia in Hybrid LLMs: When CoT Fine-Tuning Breaks Long-Range Recall, and How to Fix ItLanguage Models Need Sleep: Learning to Self-Modify and Consolidate MemoriesChain-of-Thought Fine-TuningAutomatic Post-Editing (APE)error-aware specialization objective
Recent events (1)
Dual-adapter routing system improves knowledge editing precision in LLMs
A new arXiv paper introduces a route-specialized dual-adapter architecture for knowledge editing in LLMs, separating the concerns of writing edits (edit adapter) and suppressing them when irrelevant (locality adapter). A relevance router gates which adapter is applied, addressing the locality problem in memory-assisted editing. Evaluated on CounterFact, zsRE, and MQuAKE benchmarks using Llama-3.1-8B-Instruct and Qwen3-8B, the method achieves best-in-class probability-preference accuracy across all three datasets. Ablations show the gain comes from the architectural separation rather than increased parameter capacity.