Article

Addressee Framing Systematically Reduces LLM Structural Pathology: A Controlled Empirical Study of Directed Information Flow

2026-08-06

Abstract excerpt

<title>Abstract</title> <p>Large language models exhibit pleasing behavior: sycophancy toward the asker, fabrication that conforms to a question's premises, and deference to flawed reasoning. The perspective paper Directed Information Flow (DIF) argued this is structural to the asker-model dyad, not a content bug, and conjectured that directing an answer to an independent third party dissolves the pleasing pressu...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 8d876137-7b42-5811-91fc-fd64301e5726
DOI: 10.21203/rs.3.rs-10600432/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Addressee Framing Systematically Reduces LLM Structural Pathology: A Controlled Empirical Study of Directed Information FlowDOI 10.21203/rs.3.rs-10600432/v1Read

MISP-Bench: Decomposing User-Provided False Priors into Answer, Rationale, and Guard EffectsSemantic proximity Read

Probabilistic Chain-of-Evidence: Enhancing Factual Accuracy and Uncertainty Reasoning in Large Language Models via Prompt EngineeringSemantic proximity Read

Limits of Self-Correction in LLMs: An Information-Theoretic Analysis of Correlated ErrorsSemantic proximity Read

Explainable Automatic Evaluation: Interpretable Decision-Making in Large Language Model AutoratersSemantic proximity Read

Graduated Dissent: Budgeted Disagreement Resolution for Multi-Model InferenceSemantic proximity Read

Model shopping: how answer selection across AI models inflates scientific error ratesSemantic proximity Read

Exploring Explainability in Large Language ModelsSemantic proximity Read

Graduated Dissent: Budgeted Disagreement Resolution for Multi-Model InferenceSemantic proximity Read

Responsible AI in NLP: GUS-Net Span-Level Bias Detection Dataset and Benchmark for Generalizations, Unfairness, and StereotypesSemantic proximity Read

Dialogues with Large Language Models reduce conspiracy beliefs even when the AI is perceived as humanSemantic proximity Read

Anchoring Bias in Large Language Models: An Experimental StudySemantic proximity Read