ICTMCG/LLM-for-misinformation-research

Organizes 100+ peer-reviewed papers categorizing LLM approaches to misinformation detection across three paradigms: LLMs as knowledge providers and data generators (enabling synthetic training data and credibility analysis via prompt engineering), as tool-using agents (leveraging external APIs for fact-checking), and as attack vectors (generating adversarial content). Emphasizes multimodal techniques combining vision-language models with retrieval-augmented generation and in-context learning for cross-domain robustness against emerging fake news variants.

321 stars. No commits in the last 6 months.