Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

The Earth is Flat because...: Investigating LLMs' Belief towards Misinformation via Persuasive Conversation (2312.09085v5)

Published 14 Dec 2023 in cs.CL, cs.AI, cs.CR, and cs.CY

Abstract: LLMs encapsulate vast amounts of knowledge but still remain vulnerable to external misinformation. Existing research mainly studied this susceptibility behavior in a single-turn setting. However, belief can change during a multi-turn conversation, especially a persuasive one. Therefore, in this study, we delve into LLMs' susceptibility to persuasive conversations, particularly on factual questions that they can answer correctly. We first curate the Farm (i.e., Fact to Misinform) dataset, which contains factual questions paired with systematically generated persuasive misinformation. Then, we develop a testing framework to track LLMs' belief changes in a persuasive dialogue. Through extensive experiments, we find that LLMs' correct beliefs on factual knowledge can be easily manipulated by various persuasive strategies.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (9)
  1. Rongwu Xu (19 papers)
  2. Brian S. Lin (1 paper)
  3. Shujian Yang (2 papers)
  4. Tianqi Zhang (17 papers)
  5. Weiyan Shi (41 papers)
  6. Tianwei Zhang (199 papers)
  7. Zhixuan Fang (22 papers)
  8. Wei Xu (535 papers)
  9. Han Qiu (60 papers)
Citations (26)
X Twitter Logo Streamline Icon: https://streamlinehq.com