Fact-checking is to verify the factuality of a claim by examining the available evidence. High-quality evidence plays a vital role in enhancing fact-checking systems. However, the provision of high-quality evidence for the system poses a challenge. To tackle this challenge, we propose a method based on a LLMs to automatically retrieve and summarize evidence from the Web. Furthermore, we construct RU22Fact, a multilingual explainable fact-checking dataset on the Russia-Ukraine conflict in 2022, each containing real-world claims, optimized evidence, and referenced explanation.

Exploring a method to enhanced the rumor detection ability of foundation models by small large language models.

Exploring a new paradigm for rumor detection with LLMs in tuned-free.

Text-based logical reasoning requires the model to understand the semantics of input text, and then understand the complex logical relationships within the text. Previous works generate instances based on logical expressions entailed within the input text. And we argue that external commonsense knowledge is still necessary for restoring the complete reasoning chains. To address this issue, in this paper, we propose CSKE, a commonsense knowledge enhanced text extension framework. CSKE incorporates abundant commonsense from an external knowledge base to restore the missing logical expressions and encodes more logical relationships.

This is the demonstration system we developed.