PERSUASIONTRACE introduces a Bayesian-network simulated target for multi-turn persuasion that matches human belief dynamics (81 vs 80) better than LLM baselines (64) and enables process-level evaluation.
Costello, Gordon Pennycook, and David G
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 8roles
background 1polarities
background 1representative citing papers
LOVER creates an unsupervised logic-regularized verifier that reaches 95% of supervised verifier performance on reasoning tasks across 10 datasets.
Self-play LLM trajectories form model-specific attractors that asymmetrically influence mixed-play partners' stylistic choices and stances across 7 models and 20 topics.
Survival analysis of three years of X posts shows conspiracy claims with greater semantic mutations have substantially longer lifespans, linked to changes in pronouns, social words, cognitive terms, and actor-action-target structures.
A 2x2 between-subjects experiment finds contextualization lowers AI persuasiveness but warmth restores it through crossover interaction, with reliance invariant to design, trust predicting outcomes independently, and AI literacy decoupling trust from behavior.
Attitude-congruent AI dialogues reduce immediate affective and opinion polarization more than incongruent ones, while incongruent dialogues increase cognitive trait empathy over two weeks.
Authors propose a four-stage framework to analyze opportunities and risks of generative AI across the health information journey from public sources to clinical care.
A survey-experiment with 236 participants shows most believe myths about gig worker vulnerabilities and that targeted counterarguments can reduce those beliefs.
citing papers explorer
-
A Model of Multi-turn Human Persuadability Using Probabilistic Belief Tracing
PERSUASIONTRACE introduces a Bayesian-network simulated target for multi-turn persuasion that matches human belief dynamics (81 vs 80) better than LLM baselines (64) and enables process-level evaluation.
-
Logic-Regularized Verifier Elicits Reasoning from LLMs
LOVER creates an unsupervised logic-regularized verifier that reaches 95% of supervised verifier performance on reasoning tasks across 10 datasets.
-
Attractor States Emerge in Multi-Turn LLM Conversations
Self-play LLM trajectories form model-specific attractors that asymmetrically influence mixed-play partners' stylistic choices and stances across 7 models and 20 topics.
-
Language Mutations Sustain the Persistences of Conspiracy Theories on Social Media
Survival analysis of three years of X posts shows conspiracy claims with greater semantic mutations have substantially longer lifespans, linked to changes in pronouns, social words, cognitive terms, and actor-action-target structures.
-
Personalized to Persuade: The Effects of Contextualization and Warmth on Trust and Reliance in Conversational AI
A 2x2 between-subjects experiment finds contextualization lowers AI persuasiveness but warmth restores it through crossover interaction, with reliance invariant to design, trust predicting outcomes independently, and AI literacy decoupling trust from behavior.
-
Divergent Paths to Depolarization: Dialogue Design Determines the Prosocial Benefits of AI-Assisted Political Argumentation
Attitude-congruent AI dialogues reduce immediate affective and opinion polarization more than incongruent ones, while incongruent dialogues increase cognitive trait empathy over two weeks.
-
Opportunities and Risks of Generative AI through the Health Information Journey
Authors propose a four-stage framework to analyze opportunities and risks of generative AI across the health information journey from public sources to clinical care.
-
Understanding, Challenging, and Demystifying Perceptions of Gig Worker Vulnerabilities
A survey-experiment with 236 participants shows most believe myths about gig worker vulnerabilities and that targeted counterarguments can reduce those beliefs.