SCAN: Sparse Circuit Anchor Interpretable Neuron for Lifelong Knowledge Editing
This addresses the problem of model collapse during lifelong knowledge editing for LLM users, though it appears incremental as it builds on existing sparse editing approaches.
The paper tackles catastrophic forgetting in large language models during sequential knowledge editing by proposing SCAN, a sparse editing framework that maintains model integrity on benchmarks like MMLU and GSM8K even after 3,000 edits, while other methods deteriorate and collapse.
Large Language Models (LLMs) often suffer from catastrophic forgetting and collapse during sequential knowledge editing. This vulnerability stems from the prevailing dense editing paradigm, which treats models as black boxes and relies on coarse-grained parameter interventions that inevitably disrupt preserved knowledge. To address this, we propose SCAN (a sparse editing framework based on Sparse Circuit Anchored Neuron) which transforms editing into a mechanism-aware manipulation by constructing a knowledge circuit via Sparse Transcoders. Experiments on Gemma2, Qwen3, and Llama3.1 across CounterFact, ZsRE and WikiFactDiff demonstrate that SCAN achieves a superior performance, maintaining model integrity on benchmarks like MMLU and GSM8K even after 3,000 sequential edits, whereas other existing methods deteriorate progressively as editing accumulates, eventually resulting in model collapse.