# FABLE: Fine-grained Fact Anchoring for Unstructured Model Editing Source: [FABLE: Fine-grained Fact Anchoring for Unstructured Model Editing]() Collection index: [llms.txt]() Data updated: 2026-04-16T05:00:11.996Z Authors: Peng Wang, Biyu Zhou, Xuehai Tang, Jizhong Han, Songlin Hu Published: 2026-04-14T10:41:36.000Z ## Abstract Unstructured model editing aims to update models with real-world text, yet existing methods often memorize text holistically without reliable fine-grained fact access. To address this, we propose FABLE, a hierarchical framework that decouples fine-grained fact injection from holistic text generation. FABLE follows a two-stage, fact-first strategy: discrete facts are anchored in shallow layers, followed by minimal updates to deeper layers to produce coherent text. This decoupling resolves the mismatch between holistic recall and fine-grained fact access, reflecting the unidirectional Transformer flow in which surface-form generation amplifies rather than corrects underlying fact representations. We also introduce UnFine, a diagnostic benchmark with fine-grained question-answer pairs and fact-level metrics for systematic evaluation. Experiments show that FABLE substantially improves fine-grained question answering while maintaining state-of-the-art holistic editing performance. Our code is publicly available at https://github.com/caskcsg/FABLE. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Fine-grained Knowledge Recall on UnFine-UnKE](): Fine-grained Knowledge Recall; UnFine-UnKE; Bert-Score: 65.63. - [Fine-grained Knowledge Recall on UnFine-CF](): Fine-grained Knowledge Recall; UnFine-CF; Bert-Score: 71.89. - [Fine-grained Knowledge Recall on UnFine-MQ](): Fine-grained Knowledge Recall; UnFine-MQ; Bert-Score: 66.99. - [Holistic Unstructured Text Generation on UnFine-UnKE](): Holistic Unstructured Text Generation; UnFine-UnKE; Bert-Score: 99.36. - [Holistic Unstructured Text Generation on UnFine-MQ](): Holistic Unstructured Text Generation; UnFine-MQ; Bert-Score: 98.84. - [Holistic Unstructured Text Generation on UnFine-CF](): Holistic Unstructured Text Generation; UnFine-CF; Bert-Score: 99.28. ## Original sources and implementation [Original paper]() [PDF]()