A Model of Multi-turn Human Persuadability Using Probabilistic Belief Tracing
Quick Answer
PERSUASIONTRACE introduces a framework for analyzing multi-turn persuasion in human-LLM interactions, revealing that LLMs effectively influence beliefs across various topics.
Quick Take
A Bayesian-network simulated target closely mimics human belief dynamics, scoring 81 compared to 64 for baseline , enhancing the evaluation of persuasive systems beyond endpoint measures.
Key Points
- PERSUASIONTRACE records multi-turn belief updates and annotates rhetorical strategies.
- Human targets cluster into two groups based on belief update patterns.
- LLMs demonstrate persuasive abilities in generic and personalized contexts.
- Bayesian-network targets achieve human-like belief dynamics with an 81 score.
- Traditional LLM simulators fail to replicate realistic human belief changes.
Paper Resources
Source Excerpt
arXiv:2606. 05330v1 Announce Type: new Abstract: can shift human beliefs across high-stakes domains, but most persuasion studies rely on pre/post belief change. These endpoint measures identify whether persuasion occurred, yet miss where and how beliefs moved within a dialogue. We present PERSUASIONTRACE, a framework for studying persuasion in human-LLM interaction. …
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →RF-Agent: A Practical Framework for Building Language Agents for RFIC Design
RF-Agent introduces a novel framework for RF circuit design using , creating a unique RF-domain reasoning dataset with over 11,000 samples. The study reveals that domain-specific supervised fine-tuning and semantic retrieval strategies significantly enhance RF reasoning performance, particularly for smaller models.