從機制層面理解為何記憶化知識在大型語言模型微調中無法泛化

摘要

微調大型語言模型以注入新知識面臨一個關鍵挑戰：模型能快速記憶新事實，卻無法將這些知識用於下游推理任務。我們將此失敗情形形式化為「**知用鴻溝**」，其特徵在於記憶與泛化之間存在準確率差距及時間延遲。為理解此現象，我們以未見知識微調大型語言模型，並運用名為「自我修補」的新型干預技術，監控知識在模型內部的空間滲透動態。自我修補能辨識出透過重新定位表徵可顯著改善失敗泛化案例的活化位置。這些結果與「知識電路錯位假說」一致：記憶的表徵可存在於模型內部，但可能未被導向至計算有效的層級。為驗證此診斷發現的實用性，我們設計一套簡單啟發式策略，可在泛化失敗案例中回復 58–75% 的理論上限改善空間。實驗跨領域進行以確保此發現的穩健性。

English

Fine-tuning LLMs to inject new knowledge faces a critical challenge: LLMs can quickly memorize new facts, yet fail to use them for downstream reasoning tasks. We formalize this failure as the \textbf{Knowing--Using Gap}, characterized by an accuracy gap and a temporal lag between memorization and generalization. To understand this phenomenon, we fine-tune LLMs with unseen knowledge and monitor the spatial permeation dynamics of the knowledge internally using a novel intervention technique called self-patching. Self-patching identifies activation locations where relocating representations substantially improves failed generalization cases. These results are consistent with a knowledge-circuit misalignment hypothesis: memorized representations can exist internally but may not be routed to computation-effective layers. To demonstrate the practicality of this diagnostic finding, we design a simple heuristic strategy which recovers 58--75\% of the oracle headroom in generalization failure. Experiments are done cross-domain for the robustness of this finding.