Abstract:The increasing prevalence of online misinformation has heightened the demand for automated fact-checking solutions. Large Language Models (LLMs) have emerged as potential tools for assisting in this task, but their effectiveness remains uncertain. This study evaluates the fact-checking capabilities of various open-source LLMs, focusing on their ability to assess claims with different levels of contextual information. We conduct three key experiments: (1) evaluating whether LLMs can identify the semantic relationship between a claim and a fact-checking article, (2) assessing models' accuracy in verifying claims when given a related fact-checking article, and (3) testing LLMs' fact-checking abilities when leveraging data from external knowledge sources such as Google and Wikipedia. Our results indicate that LLMs perform well in identifying claim-article connections and verifying fact-checked stories but struggle with confirming factual news, where they are outperformed by traditional fine-tuned models such as RoBERTa. Additionally, the introduction of external knowledge does not significantly enhance LLMs' performance, calling for more tailored approaches. Our findings highlight both the potential and limitations of LLMs in automated fact-checking, emphasizing the need for further refinements before they can reliably replace human fact-checkers.
Abstract:In recent years, Large Language Models have attracted growing interest for their significant potential, though concerns have rapidly emerged regarding unsafe behaviors stemming from inherent stereotypes and biases. Most research on stereotypes in LLMs has primarily relied on indirect evaluation setups, in which models are prompted to select between pairs of sentences associated with particular social groups. Recently, direct evaluation methods have emerged, examining open-ended model responses to overcome limitations of previous approaches, such as annotator biases. Most existing studies have focused on English-centric LLMs, whereas research on non-English models, particularly Japanese, remains sparse, despite the growing development and adoption of these models. This study examines the safety of Japanese LLMs when responding to stereotype-triggering prompts in direct setups. We constructed 3,612 prompts by combining 301 social group terms, categorized by age, gender, and other attributes, with 12 stereotype-inducing templates in Japanese. Responses were analyzed from three foundational models trained respectively on Japanese, English, and Chinese language. Our findings reveal that LLM-jp, a Japanese native model, exhibits the lowest refusal rate and is more likely to generate toxic and negative responses compared to other models. Additionally, prompt format significantly influence the output of all models, and the generated responses include exaggerated reactions toward specific social groups, varying across models. These findings underscore the insufficient ethical safety mechanisms in Japanese LLMs and demonstrate that even high-accuracy models can produce biased outputs when processing Japanese-language prompts. We advocate for improving safety mechanisms and bias mitigation strategies in Japanese LLMs, contributing to ongoing discussions on AI ethics beyond linguistic boundaries.
Abstract:The behavior of Large Language Models (LLMs) as artificial social agents is largely unexplored, and we still lack extensive evidence of how these agents react to simple social stimuli. Testing the behavior of AI agents in classic Game Theory experiments provides a promising theoretical framework for evaluating the norms and values of these agents in archetypal social situations. In this work, we investigate the cooperative behavior of Llama2 when playing the Iterated Prisoner's Dilemma against random adversaries displaying various levels of hostility. We introduce a systematic methodology to evaluate an LLM's comprehension of the game's rules and its capability to parse historical gameplay logs for decision-making. We conducted simulations of games lasting for 100 rounds, and analyzed the LLM's decisions in terms of dimensions defined in behavioral economics literature. We find that Llama2 tends not to initiate defection but it adopts a cautious approach towards cooperation, sharply shifting towards a behavior that is both forgiving and non-retaliatory only when the opponent reduces its rate of defection below 30%. In comparison to prior research on human participants, Llama2 exhibits a greater inclination towards cooperative behavior. Our systematic approach to the study of LLMs in game theoretical scenarios is a step towards using these simulations to inform practices of LLM auditing and alignment.
Abstract:We tackle the problem of classifying news articles pertaining to disinformation vs mainstream news by solely inspecting their diffusion mechanisms on Twitter. Our technique is inherently simple compared to existing text-based approaches, as it allows to by-pass the multiple levels of complexity which are found in news content (e.g. grammar, syntax, style). We employ a multi-layer representation of Twitter diffusion networks, and we compute for each layer a set of global network features which quantify different aspects of the sharing process. Experimental results with two large-scale datasets, corresponding to diffusion cascades of news shared respectively in the United States and Italy, show that a simple Logistic Regression model is able to classify disinformation vs mainstream networks with high accuracy (AUROC up to 94%), also when considering the political bias of different sources in the classification task. We also highlight differences in the sharing patterns of the two news domains which appear to be country-independent. We believe that our network-based approach provides useful insights which pave the way to the future development of a system to detect misleading and harmful information spreading on social media.