The performance of large language models (LLMs) has recently improved to the point where the models can perform well on many language tasks. We show here that for the first time, the models can also generate coherent and valid formal analyses of linguistic data and illustrate the vast potential of large language models for analyses of their metalinguistic abilities. LLMs are primarily trained on language data in the form of text; analyzing and evaluating their metalinguistic abilities improves our understanding of their general capabilities and sheds new light on theoretical models in linguistics. In this paper, we probe into GPT-4's metalinguistic capabilities by focusing on three subfields of formal linguistics: syntax, phonology, and semantics. We outline a research program for metalinguistic analyses of large language models, propose experimental designs, provide general guidelines, discuss limitations, and offer future directions for this line of research. This line of inquiry also exemplifies behavioral interpretability of deep learning, where models' representations are accessed by explicit prompting rather than internal representations.

Large language models (LLMs) have recently gained popularity. However, the impact of their general availability through ChatGPT on sensitive areas of everyday life, such as education, remains unclear. Nevertheless, the societal impact on established educational methods is already being experienced by both students and educators. Our work focuses on higher physics education and examines problem solving strategies. In a study, students with a background in physics were assigned to solve physics exercises, with one group having access to an internet search engine (N=12) and the other group being allowed to use ChatGPT (N=27). We evaluated their performance, strategies, and interaction with the provided tools. Our results showed that nearly half of the solutions provided with the support of ChatGPT were mistakenly assumed to be correct by the students, indicating that they overly trusted ChatGPT even in their field of expertise. Likewise, in 42% of cases, students used copy & paste to query ChatGPT -- an approach only used in 4% of search engine queries -- highlighting the stark differences in interaction behavior between the groups and indicating limited reflection when using ChatGPT. In our work, we demonstrated a need to (1) guide students on how to interact with LLMs and (2) create awareness of potential shortcomings for users.

Large Language Models (LLMs), such as ChatGPT, are becoming increasingly sophisticated, demonstrating capabilities that closely resemble those of humans. These AI models are playing an essential role in assisting humans with a wide array of tasks in daily life. A significant application of AI is its use as a chat agent, responding to human inquiries across various domains. Current LLMs have shown proficiency in answering general questions. However, basic question-answering dialogue often falls short in complex diagnostic scenarios, such as legal or medical consultations. These scenarios typically necessitate Task-Oriented Dialogue (TOD), wherein an AI chat agent needs to proactively pose questions and guide users towards specific task completion. Previous fine-tuning models have underperformed in TOD, and current LLMs do not inherently possess this capability. In this paper, we introduce DiagGPT (Dialogue in Diagnosis GPT), an innovative method that extends LLMs to TOD scenarios. Our experiments reveal that DiagGPT exhibits outstanding performance in conducting TOD with users, demonstrating its potential for practical applications.

This paper presents a comprehensive survey of ChatGPT-related (GPT-3.5 and GPT-4) research, state-of-the-art large language models (LLM) from the GPT series, and their prospective applications across diverse domains. Indeed, key innovations such as large-scale pre-training that captures knowledge across the entire world wide web, instruction fine-tuning and Reinforcement Learning from Human Feedback (RLHF) have played significant roles in enhancing LLMs' adaptability and performance. We performed an in-depth analysis of 194 relevant papers on arXiv, encompassing trend analysis, word cloud representation, and distribution analysis across various application domains. The findings reveal a significant and increasing interest in ChatGPT-related research, predominantly centered on direct natural language processing applications, while also demonstrating considerable potential in areas ranging from education and history to mathematics, medicine, and physics. This study endeavors to furnish insights into ChatGPT's capabilities, potential implications, ethical concerns, and offer direction for future advancements in this field.

After the inception of emotion recognition or affective computing, it has increasingly become an active research topic due to its broad applications. Over the past couple of decades, emotion recognition models have gradually migrated from statistically shallow models to neural network-based deep models, which can significantly boost the performance of emotion recognition models and consistently achieve the best results on different benchmarks. Therefore, in recent years, deep models have always been considered the first option for emotion recognition. However, the debut of large language models (LLMs), such as ChatGPT, has remarkably astonished the world due to their emerged capabilities of zero/few-shot learning, in-context learning, chain-of-thought, and others that are never shown in previous deep models. In the present paper, we comprehensively investigate how the LLMs perform in emotion recognition in terms of diverse aspects, including in-context learning, few-short learning, accuracy, generalisation, and explanation. Moreover, we offer some insights and pose other potential challenges, hoping to ignite broader discussions about enhancing emotion recognition in the new era of advanced and generalised large models.

Anonymity of both natural and legal persons in court rulings is a critical aspect of privacy protection in the European Union and Switzerland. With the advent of LLMs, concerns about large-scale re-identification of anonymized persons are growing. In accordance with the Federal Supreme Court of Switzerland, we explore the potential of LLMs to re-identify individuals in court rulings by constructing a proof-of-concept using actual legal data from the Swiss federal supreme court. Following the initial experiment, we constructed an anonymized Wikipedia dataset as a more rigorous testing ground to further investigate the findings. With the introduction and application of the new task of re-identifying people in texts, we also introduce new metrics to measure performance. We systematically analyze the factors that influence successful re-identifications, identifying model size, input length, and instruction tuning among the most critical determinants. Despite high re-identification rates on Wikipedia, even the best LLMs struggled with court decisions. The complexity is attributed to the lack of test datasets, the necessity for substantial training resources, and data sparsity in the information used for re-identification. In conclusion, this study demonstrates that re-identification using LLMs may not be feasible for now, but as the proof-of-concept on Wikipedia showed, it might become possible in the future. We hope that our system can help enhance the confidence in the security of anonymized decisions, thus leading to the courts being more confident to publish decisions.

String data is common in real-world datasets: 67.6% of values in a sample of 1.8 million real Excel spreadsheets from the web were represented as text. Systems that successfully clean such string data can have a significant impact on real users. While prior work has explored errors in string data, proposed approaches have often been limited to error detection or require that the user provide annotations, examples, or constraints to fix the errors. Furthermore, these systems have focused independently on syntactic errors or semantic errors in strings, but ignore that strings often contain both syntactic and semantic substrings. We introduce DataVinci, a fully unsupervised string data error detection and repair system. DataVinci learns regular-expression-based patterns that cover a majority of values in a column and reports values that do not satisfy such patterns as data errors. DataVinci can automatically derive edits to the data error based on the majority patterns and constraints learned over other columns without the need for further user interaction. To handle strings with both syntactic and semantic substrings, DataVinci uses an LLM to abstract (and re-concretize) portions of strings that are semantic prior to learning majority patterns and deriving edits. Because not all data can result in majority patterns, DataVinci leverages execution information from an existing program (which reads the target data) to identify and correct data repairs that would not otherwise be identified. DataVinci outperforms 7 baselines on both error detection and repair when evaluated on 4 existing and new benchmarks.

Modern large language models (LLMs) that rely on attention mechanisms are typically trained with fixed context lengths which enforce upper limits on the length of input sequences that they can handle at evaluation time. To use these models on sequences longer than the train-time context length, one might employ techniques from the growing family of context length extrapolation methods -- most of which focus on modifying the system of positional encodings used in the attention mechanism to indicate where tokens or activations are located in the input sequence. We conduct a wide survey of existing methods of context length extrapolation on a base LLaMA or LLaMA 2 model, and introduce some of our own design as well -- in particular, a new truncation strategy for modifying the basis for the position encoding. We test these methods using three new evaluation tasks (FreeFormQA, AlteredNumericQA, and LongChat-Lines) as well as perplexity, which we find to be less fine-grained as a measure of long context performance of LLMs. We release the three tasks publicly as datasets on HuggingFace. We discover that linear scaling is the best method for extending context length, and show that further gains can be achieved by using longer scales at evaluation time. We also discover promising extrapolation capabilities in the truncated basis. To support further research in this area, we release three new 13B parameter long-context models which we call Giraffe: 4k and 16k context models trained from base LLaMA-13B, and a 32k context model trained from base LLaMA2-13B. We also release the code to replicate our results.

The rapid expansion of the digital world has propelled sentiment analysis into a critical tool across diverse sectors such as marketing, politics, customer service, and healthcare. While there have been significant advancements in sentiment analysis for widely spoken languages, low-resource languages, such as Bangla, remain largely under-researched due to resource constraints. Furthermore, the recent unprecedented performance of Large Language Models (LLMs) in various applications highlights the need to evaluate them in the context of low-resource languages. In this study, we present a sizeable manually annotated dataset encompassing 33,605 Bangla news tweets and Facebook comments. We also investigate zero- and few-shot in-context learning with several language models, including Flan-T5, GPT-4, and Bloomz, offering a comparative analysis against fine-tuned models. Our findings suggest that monolingual transformer-based models consistently outperform other models, even in zero and few-shot scenarios. To foster continued exploration, we intend to make this dataset and our research tools publicly available to the broader research community. In the spirit of further research, we plan to make this dataset and our experimental resources publicly accessible to the wider research community.

The growing sophistication of Business Email Compromise (BEC) and spear phishing attacks poses significant challenges to organizations worldwide. The techniques featured in traditional spam and phishing detection are insufficient due to the tailored nature of modern BEC attacks as they often blend in with the regular benign traffic. Recent advances in machine learning, particularly in Natural Language Understanding (NLU), offer a promising avenue for combating such attacks but in a practical system, due to limitations such as data availability, operational costs, verdict explainability requirements or a need to robustly evolve the system, it is essential to combine multiple approaches together. We present CAPE, a comprehensive and efficient system for BEC detection that has been proven in a production environment for a period of over two years. Rather than being a single model, CAPE is a system that combines independent ML models and algorithms detecting BEC-related behaviors across various email modalities such as text, images, metadata and the email's communication context. This decomposition makes CAPE's verdicts naturally explainable. In the paper, we describe the design principles and constraints behind its architecture, as well as the challenges of model design, evaluation and adapting the system continuously through a Bayesian approach that combines limited data with domain knowledge. Furthermore, we elaborate on several specific behavioral detectors, such as those based on Transformer neural architectures.