AnyEdit++: Adaptive Long-Form Knowledge Editing via Bayesian Surprise
This work addresses the challenge of maintaining coherence when editing long-form knowledge in LLMs, a problem for practitioners needing precise and consistent model updates.
AnyEdit++ introduces a structure-aware framework for long-form knowledge editing in LLMs, using Bayesian Surprise to adaptively segment text. It achieves superior performance and robustness over state-of-the-art baselines across mathematical reasoning, code generation, and narrative tasks.
Editing complex, long-form knowledge in Large Language Models remains a significant challenge due to the difficulty of maintaining generation coherence. Existing autoregressive methods like AnyEdit alleviate length constraints but rely on Fixed-window Chunking, which disregards logical structure and compromises consistency. To address this, we present AnyEdit++, a structure-aware framework incorporating Bayes-Chunk, an adaptive segmentation mechanism that dynamically identifies semantic boundaries based on Bayesian Surprise. We underpin this approach with a theoretical framework establishing two key principles: (1) Structural Independence: we prove that cross-segment interference is minimized when anchor keys are geometrically orthogonal (a condition naturally satisfied by our surprisal-based boundaries but violated by fixed windows), and (2) Causal Locality: we demonstrate that updates injected at these semantic peaks yield strictly superior control compared to arbitrary split points. Extensive experiments across mathematical reasoning, code generation, and narrative tasks demonstrate that AnyEdit++ achieves superior performance and robustness compared to state-of-the-art baselines, validating that structural awareness is critical for effective long-form knowledge editing.