Multi-ToM: Evaluating Multilingual Theory of Mind Capabilities in Large Language Models

Theory of Mind (ToM) refers to the cognitive ability to infer and attribute mental states to oneself and others. As large language models (LLMs) are increasingly evaluated for social and cognitive capabilities, it remains unclear to what extent these models demonstrate ToM across diverse languages and cultural contexts. In this paper, we introduce a comprehensive study of multilingual ToM capabilities aimed at addressing this gap. Our approach includes two key components: (1) We translate existing ToM datasets into multiple languages, effectively creating a multilingual ToM dataset and (2) We enrich these translations with culturally specific elements to reflect the social and cognitive scenarios relevant to diverse populations. We conduct extensive evaluations of six state-of-the-art LLMs to measure their ToM performance across both the translated and culturally adapted datasets. The results highlight the influence of linguistic and cultural diversity on the models' ability to exhibit ToM, and questions their social reasoning capabilities. This work lays the groundwork for future research into enhancing LLMs' cross-cultural social cognition and contributes to the development of more culturally aware and socially intelligent AI systems. All our data and code are publicly available.

On the Identification of the Energy related Issues from the App Reviews

The energy inefficiency of the apps can be a major issue for the app users which is discussed on App Stores extensively. Previous research has shown the importance of investigating the energy related app reviews to identify the major causes or categories of energy related user feedback. However, there is no study that efficiently extracts the energy related app reviews automatically. In this paper, we empirically study different techniques for automatic extraction of the energy related user feedback. We compare the accuracy, F1-score and run time of numerous machine-learning models with relevant feature combinations and relatively modern Neural Network-based models. In total, 60 machine learning models are compared to 30 models that we build using six neural network architectures and three word embedding models. We develop a visualization tool for this study through which a developer can traverse through this large-scale result set. The results show that neural networks outperform the other machine learning techniques and can achieve the highest F1-score of 0.935. To replicate the research results, we have open sourced the interactive visualization tool. After identifying the best results and extracting the energy related reviews, we further compare various techniques to help the developers automatically investigate the emerging issues that might be responsible for energy inefficiency of the apps. We experiment the previously used string matching with results obtained from applying two of the state-of-the-art topic modeling algorithms, OBTM and AOLDA. Finally, we run a qualitative study performed in collaboration with developers and students from different institutions to determine their preferences for identifying necessary topics from previously categorized reviews, which shows OBTM produces the most helpful results.