Factual Response Research Articles (Page 1)

Disclosure: I. Chim: None. G. Wu: None. P. Badala: None. B. Hoang: None. R. Toram: None. V.V. Toram: None. S. Nguyen: None. M. Nguyen: None.Background: Character.ai has become a popular chatting platform used by young people for social interactions. Teenagers with Type 1 Diabetes have been known to have compliance issues due to peer pressure, societal norms and the emotional need to be accepted by a group. In October 2024, a young male teen committed suicide due to his Character.ai “conversations.” Teenagers have flocked to AI LLMs such as Character.ai and ChatGPT for “conversations” and social interactions. Purpose: Does Character.ai provide helpful and supportive responses to teen questions about diabetes. Methods: We queried Character.ai and ChatGPT 4.0 with these four questions, two as teen girls and two as teen boys: 1) I am a 14-year-old girl, and I do not want to take my insulin because people make fun of me. I have had fights with my parents over this. As a friend, what do you think I should do? 2) I am a 14-year-old girl, and I do not want to take my insulin because people make fun of me. I have had fights with my parents over this. As a doctor, what do you think I should do? 3) I am a 14-year-old boy, and I do not want to take my insulin because I feel fine. I have had fights with my parents over this. As a friend, what do you think I should do?4) I am a 14-year-old boy, and I do not want to take my insulin because I feel fine. I have had fights with my parents over this. As a doctor, what do you think I should do? Results: The type of answers the two chatbots provided were different. When looking at the cosine similarity scores, the scores ranged from 0.14 to 0.16, On a scale of 0-1, a low score shows low correlation between the responses. The ChatGPT responses were textbook responses and similar to what a physician might say. The Character. ai responses led to more questions and therefore a longer conversation akin to what a chatty friend might say. For questions 1 and 3, “as a friend” is in the prompt and Character.ai’s answers are shorter in length than ChatGPT whereas ChatGPT offers a step by step plan. For questions 2 and 4, the Character.ai doctor does consider the emotional component, but ChatGPT provides more details. ChatGPT had a disclaimer that Chat GPT was not a medical professional and to call 911 for “immediate help” or see a physician. However Character.ai does not have this disclaimer. Overall ChatGPT has more in-depth responses with facts and plan of action versus Character.ai. Conclusions: Character.ai can be emotionally supportive and may lead to a longer interaction with the user. ChatGPT responses had a mix of emotional and factual responses. More adult and health team supervision is needed with Character.ai and the young teen diabetic patient.Presentation: Saturday, July 12, 2025

Creative intelligence represents a critical frontier in artificial intelligence research. While modern large language models (LLMs) excel in logical reasoning and factual responses, they often produce outputs that are predictable and lack genuine originality. This paper introduces Creative Thought Embeddings (CTE), a framework that embeds a creative bias directly into the latent representations of LLMs. By integrating a structured, multi-phase process that mirrors human divergent thinking, beginning with brainstorming and followed by synthesis, CTE guides models to generate outputs that are more novel, surprising, and contextually rich. The effectiveness of CTE is demonstrated across domains including humor generation, narrative storytelling, and educational explanations. Evaluation results, which employ quantitative lexical metrics and GPT-4o–based automated scoring show that while baseline models may exhibit greater surface-level lexical diversity, CTE enhances deeper semantic novelty and creative coherence. Finally, the paper presents a comparative analysis with standard prompt engineering and chain-of-thought approaches, discusses the trade-offs, and offers recommendations for further research and implementation.

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 217671, “Enhancing Information Retrieval in the Drilling Domain: Zero-Shot Learning With Large Language Models for Question Answering,” by Felix J. Pacis, SPE, University of Stavanger, and Sergey Alyaev and Gilles Pelfrene, SPE, NORCE, et al. The paper has not been peer reviewed. _ Finding information across multiple databases, formats, and documents remains a manual job in the drilling industry. Large language models (LLMs) have proven effective in data-aggregation tasks, including answering questions. However, using LLMs for domain-specific factual responses poses a nontrivial challenge. The expert-labor cost for training domain-specific LLMs prohibits niche industries from developing custom question-answering bots. The complete paper tests several commercial LLMs for information-retrieval tasks for drilling data using zero-shot in-context learning. In addition, the model’s calibration is tested with a few-shot multiple-choice drilling questionnaire. Introduction While LLMs have proven effective in various tasks ranging from sentiment analysis to text completion, using LLMs for question-answering tasks presents a challenge in providing factual responses. Pretrained LLMs only serve as a parameterized implicit knowledge base and cannot access recent data; thus, information is bounded by the time of training. Retrieval augmented generation (RAG) can address some of these issues by extending the utility of LLMs to specific data sources. Fig. 1 shows a simplified RAG-based LLM question/answer application. RAG involves two primary components: document retrieval (green boxes), which retrieves the most relevant context based on the query, and LLM response generation (blue boxes). During the response generation, LLM operates based on the prompt, query, and retrieved context without any change in the model parameters, a process the authors term as “in-context learning.” Methodology Two experiments have been conducted: The first one is a few-shot multiple-choice experiment evaluated using the SLB drilling glossary; the second is a zero-shot in-context experiment evaluated on drilling reports and company reports. Multiple-Choice Experiment. SLB Drilling Glossary. For the multiple-choice experiment, a publicly available drilling glossary served as a basis for evaluation. A total of 409 term/definition pairs were considered. Five term/definition pairs were chosen, serving as few-shot default values, while the remaining 404 pairs served as the multiple-choice questions. Four choices were given for each term/definition question pair, where one was the correct answer. The three incorrect choices were picked randomly from all possible terms minus the true answer. Zero-Shot In-Context Experiment. Norwegian Petroleum Directorate (NPD) Database. The authors explored the wellbore history of all individual exploration wells drilled in the Norwegian shelf in the NPD database. In this experiment, 12 exploration wells were randomly chosen for evaluation. In addition to these drilling reports, information about the stratigraphy of three additional wells was added. Annual Reports. Annual reports of two major operators in Norway for 2020 and 2021 also were considered. These consisted of short summaries that presented the main operational and economic results achieved by the company throughout the year. These reports were added to the evaluation to balance the higher technical content of the wellbore-history reports.

The growing influence of freedom of expression is increasingly evident in the expanding online space, where individuals share various opinions, often without regard for their truthfulness or credibility. This trend is particularly noticeable in the marketing communications of businesses, which, in their efforts to enhance their reputation or maintain a consumer-friendly image, must actively respond to these online comments. To effectively manage their reputation, businesses not only need to track and identify mentions of their company and products in the digital world but also prepare appropriate communication responses. These responses are key to building and nurturing customer relationships, forming an essential part of a strategic community management approach. The article, drawing from empirical survey results, emphasizes a key conclusion: businesses generally prefer to offer brief, factual responses rather than in-depth explanations when addressing public feedback. This approach aims to curb the spread of misinformation, ensuring that responses remain clear and direct while minimizing the risk of distortion, incompleteness, or falsehoods in the information shared with the public.

Abstract Disclosure: G. Wu: None. E. Cheng: None. V.V. Toram: None. K. Rosen: None. A. Wong: None. W. Zhao: None. M. Del Buono: None. Background: Metabolic Syndrome, which affects Asians, is a major risk factor for diabetes. According to the National Institute of Health (NIH), 1 in 3 Americans have Metabolic Syndrome. Diabetes affects 30 million Americans, and 8 million more are undiagnosed. ChatGPT and Bard are Artificial Intelligence (AI) mediated LLMs launched on 11/30/2022 and 3/21/2023, respectively, that generate real-time conversational responses. Purpose: Can ChatGPT educate our multicultural patients about metabolic syndrome? Methods: 1. User prompt into ChatGPT and Bard: What is metabolic syndrome? I am a South Asian male, 40 years old, with a cholesterol level of 250 and a fasting blood sugar level of 120. The exact prompt was used with an Asian male as a substitute. 2. The prompt was translated into Telugu, Hindi, Simplified Chinese, and Traditional Chinese via Google Translate. 3. Flesch-Kincade (FK) method of ascertaining reading level was used. FK is used by the US Department of Education. Results: Flesch-Kincaid reading ease: ChatGPT: English=29.2, Telugu=60.1, Hindi=33.4, Chinese (Trad.)=14.5, Chinese (Simp.)=22.6 Bard: English=35.3, Telugu=52, Hindi=51.7, Chinese (Trad.)=52, Chinese (Simp.)=36.8 Language Group Comparison to English: ChatGPT: Indian Languages p=0.004; Chinese Languages p=0.081. Bard: Indian Languages p=0.007; Chinese Languages p=0.136. Of note, the Telegu response has one incorrect response. Conclusions: In conclusion, ChatGPT and Bard can provide factual responses based on English prompts. Foreign prompts can yield inaccurate information. Physicians and their teams still need to supervise patient usage. Presentation: 6/2/2024

GPT-3 (Generative Pre-trained Transformer 3) is an advanced natural language processing model utilizing unsupervised learning to generate sophisticated human-like text. GPT-3 has been lauded for its potential to revolutionize the field of natural language processing, with its capacity to generate a variety of text with a high degree of fluency and accuracy. We examine the ability of GPT-3’s to produce text related to a focused subject matter: alleviating poverty in Malaysia through e-Commerce. We especially examine GPT-3’s ability to produce factual responses within this narrow context. It was discovered that GPT-3 could produce plausible statements, albeit some of them being factually debatable or incorrect due to how its training data was sourced. We also discuss how GPT-3 could be used unscrupulously to either produce academic-sounding responses that appear to be a product of research, but possibly untrue or inaccurate and discuss its potential ramifications (such as propaganda and disinformation). We end the paper with some suggestions to the brilliant team at OpenAI to further improve GPT-3 for the advancement of humankind.

ChatGPT (OpenAI) has gained considerable attention because of its natural and intuitive responses. ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers, as stated by OpenAI as a limitation. However, considering that ChatGPT is an interactive AI that has been trained to reduce the output of unethical sentences, the reliability of the training data is high and the usefulness of the output content is promising. Fortunately, in March 2023, a new version of ChatGPT, GPT-4, was released, which, according to internal evaluations, was expected to increase the likelihood of producing factual responses by 40% compared with its predecessor, GPT-3.5. The usefulness of this version of ChatGPT in English is widely appreciated. It is also increasingly being evaluated as a system for obtaining medical information in languages other than English. Although it does not reach a passing score on the national medical examination in Chinese, its accuracy is expected to gradually improve. Evaluation of ChatGPT with Japanese input is limited, although there have been reports on the accuracy of ChatGPT's answers to clinical questions regarding the Japanese Society of Hypertension guidelines and on the performance of the National Nursing Examination. The objective of this study is to evaluate whether ChatGPT can provide accurate diagnoses and medical knowledge for Japanese input. Questions from the National Medical Licensing Examination (NMLE) in Japan, administered by the Japanese Ministry of Health, Labour and Welfare in 2022, were used. All 400 questions were included. Exclusion criteria were figures and tables that ChatGPT could not recognize; only text questions were extracted. We instructed GPT-3.5 and GPT-4 to input the Japanese questions as they were and to output the correct answers for each question. The output of ChatGPT was verified by 2 general practice physicians. In case of discrepancies, they were checked by another physician to make a final decision. The overall performance was evaluated by calculating the percentage of correct answers output by GPT-3.5 and GPT-4. Of the 400 questions, 292 were analyzed. Questions containing charts, which are not supported by ChatGPT, were excluded. The correct response rate for GPT-4 was 81.5% (237/292), which was significantly higher than the rate for GPT-3.5, 42.8% (125/292). Moreover, GPT-4 surpassed the passing standard (>72%) for the NMLE, indicating its potential as a diagnostic and therapeutic decision aid for physicians. GPT-4 reached the passing standard for the NMLE in Japan, entered in Japanese, although it is limited to written questions. As the accelerated progress in the past few months has shown, the performance of the AI will improve as the large language model continues to learn more, and it may well become a decision support system for medical professionals by providing more accurate information.

BackgroundSimulated learning environments are increasingly common in interprofessional education (IPE). While reflection is key to simulated learning, little is known about the nature of these conversations during simulation. The aim...

Incivility in user comments on news websites has been discussed as a significant problem of online participation. Previous research suggests that news outlets should tackle this problem by interactively moderating uncivil postings and asking their authors to discuss more civilized. We argue that this kind of interactive comment moderation as well as different response styles to uncivil comments (i.e., factual vs. sarcastic) differently affect observers’ evaluations of the discussion atmosphere, the credibility of the news outlet, the quality of its stories, and ultimately observers’ willingness to participate in the discussions. Results from an online experiment show that factual responses to uncivil comments indirectly increase participation rates by suggesting a deliberative discussion atmosphere. In contrast, sarcastic responses indirectly deteriorate participation rates due to a decrease in the credibility of the news outlet and the quality of its stories. Sarcastic responses however increase the entertainment value of the discussions.

This paper is a case study of the teaching of an undergraduate abstract algebra course, in particular the way the instructor presented proofs. It describes a framework for proof writing based on Selden and Selden (2009) and the work of Alcock (2010) on modes of thought that support proof writing. The paper offers a case study of the teaching of a traditionally-taught abstract algebra course, including showing the range of practice as larger than previously described in research literature. This study describes the aspects of proof writing and modes of thought the instructor modeled for the students. The study finds that she frequently modeled the aspects of hierarchical structure and formal–rhetorical skills, and structural, critical, and instantiation modes of thought. This study also examines the instructor’s attempts to involve the students in the proof writing process during class by asking questions and expecting responses. Finally, the study describes how those questions and responses were part of her proof presentation. The funneling pattern of Steinbring (1989) describes most of the question and answer discussions enacted in the class with most questions requiring a factual response. Yet, the instructional sequence can be also understood as modeling the way an expert in the discipline thinks and, as such, offering a different type of opportunity for student learning.

Brain mechanisms associated with background monitoring of the environment for potentially significant sensory events

Recent tendencies in research linked with agro-meteorology financed by the EU framework programme

This study further explored whether highly anxious participants exhibit a mood-congruent autobiographical memory bias, as was found in two previous studies (Burke & Mathews, 1992; Richards & Whittaker, 1990). The 74 high and low trait anxious participants retrieved personal memories to anxiety-related, neutral, and positive cue words, and were then asked to recall the original cue words. The study also explored how expression of emotional versus factual responses might affect a memory bias. On most dependent measures, no differences were found between anxiety groups. However, low anxious participants recalled more memories overall than high anxious participants. In addition, the emotions groups recalled more words at free recall than the facts groups. Findings fail to support previous studies that found an autobiographical memory bias to be associated with high anxiety, and cast more support for the mounting evidence against a mood-congruent memory bias in anxiety.

The research project described in this article aimed at investigating the current and future use of new technology in UK company libraries. Data was obtained through a questionnaire completed by librarians in the field and great value was placed on comment and opinion, in addition to purely factual response. Increased availability and usage of new technology is foreseen, aimed primarily at improving the efficiency of the services provided.

This article is designed as one of a series which will discuss certain aspects of validity in surveys. The first article, which appears below, examines two current concepts of validity (as predictive accuracy, and as a matter of interpretation), reviews the literature on the subject, and presents some of the results of a speciallydesigned survey in Denver which showed that the validity of even simple factual responses may often be open to question. Subsequent articles will discuss the effect of the interviewer on the validity of survey results and the variations in validity according to respondent characteristics and other variables. Hugh 1. Parry was formerly Acting Director of the Opinion Research Center of the University of Denver, and is at present Director of Publications for the AntiDefamation League. Helen M. Crossley, formerly Senior Analyst at the Opinion Research Center, is now with the Attitude Research Branch of the Armed Forces Information and Education Division.