Below are articles that use LLMs in their research workflows. You can use the Search option to find examples from your discipline, or for specific workflow applications you may be considering.
| Title | Type of Resource | Description of Resource | Link to Resource | Open Science | Use of LLM | Research Discipline(s) |
|---|---|---|---|---|---|---|
| Supporting Qualitative Analysis with Large Language Models: Combining Codebook with GPT-3 for Deductive Coding | Research Article | Qualitative analysis of textual contents unpacks rich and valuable information by assigning labels to the data. However, this process is often labor-intensive, particularly when working with large datasets. While recent AI-based tools demonstrate utility, researchers may not have readily available AI resources and expertise, let alone be challenged by the limited generalizability of those task-specific models. In this study, we explored the use of large language models (LLMs) in supporting deductive coding, a major category of qualitative analysis where researchers use pre-determined codebooks to label the data into a fixed set of codes. Instead of training task-specific models, a pre-trained LLM could be used directly for various tasks without fine-tuning through prompt learning. Using a curiosity-driven questions coding task as a case study, we found, by combining GPT-3 with expert-drafted codebooks, our proposed approach achieved fair to substantial agreements with expert-coded results. We lay out challenges and opportunities in using LLMs to support qualitative coding and beyond. | Qualitative Coding | Preprint | Data Analysis | Computer Science |
| Surveying Generative AI’s Economic Expectations | Research Article | I introduce a survey of economic expectations formed by querying a large language model (LLM)’s expectations of various financial and macroeconomic variables based on a sample of news articles from the Wall Street Journal between 1984 and 2021. I find the resulting expectations closely match existing surveys including the Survey of Professional Forecasters (SPF), the American Association of Individual Investors, and the Duke CFO Survey. Importantly, I document that LLM based expectations match many of the deviations from full-information rational expectations exhibited in these existing survey series. The LLM’s macroeconomic expectations exhibit under-reaction commonly found in consensus SPF forecasts. Additionally, its return expectations are extrapolative, disconnected from objective measures of expected returns, and negatively correlated with future realized returns. Finally, using a sample of articles outside of the LLM’s training period I find that the correlation with existing survey measures persists – indicating these results do not reflect memorization but generalization on the part of the LLM. My results provide evidence for the potential of LLMs to help us better understand human beliefs and navigate possible models of nonrational expectations. | arxiv.org/pdf/2305.02823.pdf | Preprint, Open Source | Data Generation, Data Analysis | Economics |
| Generative Agents: Interactive Simulacra of Human Behavior | Research Article | Believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents--computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day. To enable generative agents, we describe an architecture that extends a large language model to store a complete record of the agent's experiences using natural language, synthesize those memories over time into higher-level reflections, and retrieve them dynamically to plan behavior. We instantiate generative agents to populate an interactive sandbox environment inspired by The Sims, where end users can interact with a small town of twenty five agents using natural language. In an evaluation, these generative agents produce believable individual and emergent social behaviors: for example, starting with only a single user-specified notion that one agent wants to throw a Valentine's Day party, the agents autonomously spread invitations to the party over the next two days, make new acquaintances, ask each other out on dates to the party, and coordinate to show up for the party together at the right time. We demonstrate through ablation that the components of our agent architecture--observation, planning, and reflection--each contribute critically to the believability of agent behavior. By fusing large language models with computational, interactive agents, this work introduces architectural and interaction patterns for enabling believable simulations of human behavior. | Agents | Preprint | Data Generation | Computer Science |
| Machine vs. Human, who makes better judgment? Take Large Language Model GPT-4 For Example Machine vs. Human, who makes better judgment? Take Large Language Model GPT-4 For Example | Research Article | This essay explores the topic of human decision-making and the concept of noise, which refers to random and irrelevant factors that can affect decision-making. The essay argues that while humans are prone to noise in their decision-making processes, artificial intelligence (AI) can make less noise due to its ability to process large amounts of data and apply logical algorithms to make decisions. The essay examines examples and studies to demonstrate the impact of noise on human decision-making, including business ideas. Additionally, the essay highlights the potential that machine intuition is doing better than humans. | Machine vs Human | Preprint | Data Collection | |
| Can ChatGPT and Bard Generate Aligned Assessment Items? A Reliability Analysis against Human Performance | Research Article | ChatGPT and Bard are AI chatbots based on Large Language Models (LLM) that are slated to promise different applications in diverse areas. In education, these AI technologies have been tested for applications in assessment and teaching. In assessment, AI has long been used in automated essay scoring and automated item generation. One psychometric property that these tools must have to assist or replace humans in assessment is high reliability in terms of agreement between AI scores and human raters. In this paper, we measure the reliability of OpenAI ChatGP and Google Bard LLMs tools against experienced and trained humans in perceiving and rating the complexity of writing prompts. Intraclass correlation (ICC) as a performance metric showed that the inter-reliability of both the OpenAI ChatGPT and the Google Bard were low against the gold standard of human ratings. | Education Assessment | Preprint | Other | Education |
| Prompting GPT-3 to be Reliable | Research Article | Large language models (LLMs) show impressive abilities via few-shot prompting. Commercialized APIs such as OpenAI GPT-3 further increase their use in real-world language applications. However, the crucial problem of how to improve the reliability of GPT-3 is still under-explored. While reliability is a broad and vaguely defined term, we decompose reliability into four main facets that correspond to the existing framework of ML safety and are well-recognized to be important: generalizability, social biases, calibration, and factuality. Our core contribution is to establish simple and effective prompts that improve GPT-3’s reliability as it: 1) generalizes out-of-distribution, 2) balances demographic distribution and uses natural language instructions to reduce social biases, 3) calibrates output probabilities, and 4) updates the LLM’s factual knowledge and reasoning chains. With appropriate prompts, GPT-3 is more reliable than smaller-scale supervised models on all these facets. We release all processed datasets, evaluation scripts, and model predictions.1 Our systematic empirical study not only sheds new insights on the reliability of prompting LLMs, but more importantly, our prompting strategies can help practitioners more reliably use LLMs like GPT-3. | Prompting GPT-3 to be Reliable | Open Source, Open Code | ||
| Questions of science: chatting with ChatGPT about complex systems | Research Article, Use Case Example | We present an overview of the complex systems field using ChatGPT as a representation of the community's understanding. ChatGPT has learned language patterns and styles from a large dataset of internet texts, allowing it to provide answers that reflect common opinions, ideas, and language patterns found in the community. Our exploration covers both teaching and learning, and research topics. We recognize the value of ChatGPT as a source for the community's ideas. | Complex systems | Preprint | Data Collection | Other |
| Towards Agile Text Classifiers for Everyone | Research Article | Text-based safety classifiers are widely used for content moderation and increasingly to tune generative language model behavior - a topic of growing concern for the safety of digital assistants and chatbots. However, different policies require different classifiers, and safety policies themselves improve from iteration and adaptation. This paper introduces and evaluates methods for agile text classification, whereby classifiers are trained using small, targeted datasets that can be quickly developed for a particular policy. Experimenting with 7 datasets from three safety-related domains, comprising 15 annotation schemes, led to our key finding: prompt-tuning large language models, like PaLM 62B, with a labeled dataset of as few as 80 examples can achieve state-of-the-art performance. We argue that this enables a paradigm shift for text classification, especially for models supporting safer online discourse. Instead of collecting millions of examples to attempt to create universal safety classifiers over months or years, classifiers could be tuned using small datasets, created by individuals or small organizations, tailored for specific use cases, and iterated on and adapted in the time-span of a day. | Agile Classifiers | |||
| Is ChatGPT a Good Sentiment Analyzer? A Preliminary Study | Research Article | Recently, ChatGPT has drawn great attention from both the research community and the public. We are particularly curious about whether it can serve as a universal sentiment analyzer. To this end, in this work, we provide a preliminary evaluation of ChatGPT on the understanding of opinions, sentiments, and emotions contained in the text. Specifically, we evaluate it in four settings, including standard evaluation, polarity shift evaluation, open-domain evaluation, and sentiment inference evaluation. The above evaluation involves 18 benchmark datasets and 5 representative sentiment analysis tasks, and we compare ChatGPT with fine-tuned BERT and corresponding state-of-the-art (SOTA) models on end-task. Moreover, we also conduct human evaluation and present some qualitative case studies to gain a deep comprehension of its sentiment analysis capabilities. | Preprint link | Preprint | Data Science |