Skip to content
AI News HubLIVE
Original source3 min read

Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule

Summary

The paper argues that retrieval-augmented language models cannot resolve conflicts between retrieved context and parametric knowledge with any fixed rule. A decision-theoretic model shows fixed-rule performance is linear in the share of reliable contexts while the optimal score is convex, so every fixed rule is beaten somewhere. Resolving conflicts requires separate evidence about each source plus a prior on how often the context is correct. On GRACE, baselines are strong only on right or wrong contexts, never both. The proposed system—calibrated notes, two-voice reading, and a selector applying an exact prior shift—beats the best baseline at every mixture by up to 26 points, scoring 45.3 and 35.0 versus 29.4 and 27.5. It also beats CK-PLUG and TCR on ClashEval by 12 and 26 points, with g…

SourcearXiv AIAuthor: Minh Pham Van, Cong Tran
Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 26 Oct 2026]

Title:Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule

View a PDF of the paper titled Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule, by Minh Pham Van and 1 other authors

View PDF HTML (experimental)

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)

Full-text links:

Access Paper:

View a PDF of the paper titled Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule, by Minh Pham Van and 1 other authors

View PDF

HTML (experimental)

TeX Source

view license

Additional Features

Audio Summary

Current browse context:

cs.AI

new | recent | 2026-10

Change to browse by:

cs cs.CL

References & Citations

SciX

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Fixed conflict-resolution rules are provably suboptimal across different context reliability mixes.
  • Symmetric divergence between context and model knowledge carries no signal about which source is correct.
  • The method separates per-source evidence and applies an explicit prior over context correctness.
  • It outperforms the best baseline on every GRACE mixture and leads CK-PLUG and TCR on ClashEval.

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models