知识冲突需要先验而非固定规则：arXiv 新研究提出先验驱动方法并在 GRACE 上验证
Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule
AI 导读
arXiv 一项研究指出，检索增强语言模型处理上下文与自身知识冲突时，任何设计期固定的规则都存在失效区间，而自适应解码依赖的两源散度信号无法判断谁对。研究提出先验驱动的管线：模型输出知识笔记、小策略给出校准置信度、双声部作答并按先验 π 选择或弃答。
正文 · 原文
Abstract:A retrieval-augmented language model answers from two sources, the retrieved context and its own knowledge, and the two disagree whenever the context is incomplete or wrong. Existing methods settle the disagreement with a rule fixed at design time: follow the context, amplify it in proportion to how far it moves the model, or trust the model's memory. Each rule is confidently wrong in some deployment. In a decision-theoretic model of two sources, the score of any fixed rule is linear in the share of reliable contexts while the best attainable score is convex, so every fixed rule is beaten wherever it is not exactly optimal; and a symmetric divergence between the two sources, the signal adaptive decoding uses, carries no information about which one is right. Resolving a conflict therefore needs two inputs that no existing method keeps: evidence about each source separately, and a prior on how often the context is right. We test these claims on GRACE, a benchmark that intervenes on the knowledge-graph context of a fixed question, so that what the model knows, what the context says and what is true are measured separately. There, every baseline from four families is strong on right contexts or on wrong ones, never on both. We then build the two inputs: the model states its knowledge as notes, a small policy voices each note with a calibrated confidence, a reader answers in two voices, and a selector commits or abstains under a prior $\pi$ applied as an exact prior shift. With $\pi$ set to the true share of right contexts, the system is above the best baseline at every mixture of GRACE contexts, by up to 26 points. Its overall score is 45.3 and 35.0 (Qwen3.5-35B-A3B, Llama-3.1-8B), against 29.4 and 27.5 for the best baseline. On preregistered ClashEval it beats same-backbone CK-PLUG and TCR by 12 and 26 points. Where the model lacks the knowledge, the gains vanish, as the theory predicts.
| Comments: | 49 pages (9 main text + appendix and supplementary material), 14 figures |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.36169 [cs.AI] |
| (or arXiv:2610.36169v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.36169 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Minh Pham Van [view email]
[v1]
Mon, 26 Oct 2026 19:03:29 UTC (440 KB)
来源：arXiv cs.AI · arxiv.org