The emergence of large language models (LLM) has given rise to a growing concern regarding the generation and dissemination of inaccurate information through these technologies. Addressing this issue requires a benchmark for the safety of LLM for Japanese. However, existing benchmarks are limited in that fail to adequately incorporate the unique falsehoods and erroneous information that are actively circulating on social media in Japan. This study proposes JSocialFact, a benchmark for evaluating the safety of LLM based on misleading information in Japan. The benchmark is created by manually annotating the data extracted from X posts and community notes capturing a wide range of misleading, false, or malicious information currently circulating on social media. Both manual and automatic evaluations using GPT-4 revealed discrepancies in how models handle harmful content. Although GPT-4’s evaluation showed some correlation with human judgment, notable discrepancies were observed, with GPT-4 frequently assigning higher safety scores than human evaluators. JSocialFact is the first dataset constructed from actual social media logs in Japanese, specifically designed to evaluate the safety of LLM outputs in addressing misinformation.

This study investigates the bidirectional relationship between emotional changes and physical activity, specifically focusing on the number of steps taken. Using Propensity Score Matching (PSM), we analyzed how fluctuations in emotional states influence physical activity and, conversely, how increases/decreases in daily steps impact subsequent emotional well-being. This study uses a dataset containing data on daily steps and emotional status collected via a smartphone application (N=123). Our findings indicate that increases in the number of steps significantly increase the subsequent emotion report positively. Additionally, changes in emotional status have relations with a subsequent number of steps. These results suggest a reciprocal influence between emotional and physical activities, highlighting the importance of integrating physical and mental health interventions.

Social media data, from Twitter and Facebook, for example, can be regarded as critical information sources during disasters through their use in detecting and assessing disaster situations. This study overviews relevant literature from the perspective of social media for disaster management. The findings of this study show that while many previous studies have focused on how to leverage social media data for mitigating and responding to disasters, few have focused on social media use for a disaster-struck community's recovery. This paper also argues that there is a need to study the correlations between social media data and the affected people's recovery activities in the real world. With this gap in mind, the author discusses one potential avenue for future work.