[pt] Neste trabalho, apresentamos um estudo sobre o fine-tuning de um LLM (Modelo de Linguagem Amplo ou Large Language Model) pré-treinado para a sumarização abstrativa de textos longos em português. Para isso, construímos um corpus contendo uma coleção de 7.450 artigos científicos na área de Ciências da Saúde em português. Utilizamos esse corpus para o fine-tuning do modelo BERT pré-treinado para o português brasileiro (BERTimbau). Em condições semelhantes, também treinamos um segundo modelo baseado em Memória de Longo Prazo e Recorrência (LSTM) do zero, para fins de comparação. Nossa avaliação mostrou que o modelo ajustado obteve pontuações ROUGE mais altas, superando o modelo baseado em LSTM em 30 pontos no F1-score. O fine-tuning do modelo pré-treinado também se destaca em uma avaliação qualitativa feita por avaliadores a ponto de gerar a percepção de que os resumos gerados poderiam ter sido criados por humanos em uma coleção de documentos específicos do domínio das Ciências da Saúde. / [en] In this work, we present a study on the fine-tuning of a pre-trained Large Language Model for abstractive summarization of long texts in Portuguese. To do so, we built a corpus gathering a collection of 7,450 public Health Sciences papers in Portuguese. We fine-tuned a pre-trained BERT model for Brazilian Portuguese (the BERTimbau) with this corpus. In a similar condition, we also trained a second model based on Long Short-Term Memory (LSTM) from scratch for comparison purposes. Our evaluation showed that the fine-tuned model achieved higher ROUGE scores, outperforming the LSTM based by 30 points for F1-score. The fine-tuning of the pre-trained model also stands out in a qualitative evaluation performed by assessors, to the point of generating the perception that the generated summaries could have been created by humans in a specific collection of documents in the Health Sciences domain.

This thesis looks at the percentage of hallucinations in two large language models (LLM), ChatGPT 3.5 and ChatGPT 4 output for a set of prompts. This work was motivated by two factors: the release of ChatGPT 4 and its parent company OpenAI, claiming it to be much more potent than its predecessor ChatGPT 3.5, which raised questions regarding the capabilities of the LLM. Furthermore, the other factor is that ChatGPT 3.5 showcased hallucinations (creating material that is factually wrong, deceptive, or untrue.) in response to different prompts, as shown by other studies. The intended audience was members of the computer science community, such as researchers, software developers, and policymakers. The aim was to highlight large language models' potential capabilities and provide insights into their dependability. This study used a quasi-experimental study design and a systematic literature review.Our hypothesis predicted that the percentage of hallucinations (creating factually wrong, deceptive, or untrue material) would be more prevalent in ChatGPT 3.5 compared to ChatGPT 4. We based our prediction on the fact that OpenAI trained ChatGPT 4 on more material than ChatGPT 3.5. We experimented on both LLMS, and our findings supported The hypothesis. Furthermore, we looked into the literature and found studies that also agree that ChatGPT 4 is better than ChatGPT 3.5. The research concluded with suggestions for future work, like using extensive datasets and comparing the performance of different models, not only ChatGPT 3.5 and ChatGPT 4.

The release of ChatGPT has really set new standards for what an artificial intelligence chatbot should be. It has even shown its potential in answering university-level exam questions from different subjects. This research is focused on evaluating its capabilities in programming subjects. To achieve this, coding questions taken from software engineering exams were posed to the AI (N = 23) through an experiment. Then, statistical analysis was done to find out how good of a student ChatGPT is by analyzing its answer’s correctness, degree of completion, diversity of response, speed of response, extraneity, number of errors, length of response and confidence levels. GPT-3.5 is the version analyzed. The experiment was done using questions from three different programming subjects. Afterwards, results showed a 93% rate of correct answer generation, demonstrating its competence. However, it was found that the AI occasionally produces unnecessary lines of code that were not asked for and thus treated as extraneity. The confidence levels given by ChatGPT, which were always high, also didn't always align with response quality which showed the subjectiveness of the AI’s self-assessment. Answer diversity was also a concern, where most answers were repeatedly written nearly the same way. Moreover, when there was diversity in the answers, it also caused much more extraneous code. If ChatGPT was to be blind tested for a software engineering exam containing a good number of coding questions, unnecessary lines of code and comments could be what gives it away as being an AI. Nonetheless, ChatGPT was found to have great potential as a learning tool. It can offer explanations, debugging help, and coding guidance just as any other tool or person could. It is not perfect though, so it should be used with caution.

This dissertation explores the use of natural language processing (NLP) and large language models (LLMs) to analyze student self-regulated learning (SRL) strategies in response to exam wrappers. Exam wrappers are structured reflection activities that prompt students to practice SRL after they get their graded exams back. The dissertation consists of three manuscripts that compare traditional qualitative analysis with NLP-assisted approaches using transformer-based models including GPT-3.5, a state-of-the-art LLM. The data set comprises 3,800 student responses from an engineering physics course. The first manuscript develops two NLP-assisted codebooks for identifying learning strategies related to SRL in exam wrapper responses and evaluates the agreement between them and traditional qualitative analysis. The second manuscript applies a novel NLP technique called zero-shot learning (ZSL) to classify student responses into the codes developed in the first manuscript and assesses the accuracy of this method by evaluating a subset of the full dataset. The third manuscript identifies the distribution and differences of learning strategies and SRL constructs among students of different exam performance profiles using the results from the second manuscript. The dissertation demonstrates the potential of NLP and LLMs to enhance qualitative research by providing scalable, robust, and efficient methods for analyzing large corpora of textual data. The dissertation also contributes to the understanding of SRL in engineering education by revealing the common learning strategies, impediments, and SRL constructs that students report they use while preparing for exams in a first-year engineering physics course. The dissertation suggests implications, limitations, and directions for future research on NLP, LLMs, and SRL. / Doctor of Philosophy / This dissertation is about using artificial intelligence (AI) to help researchers and teachers understand how students learn from their exams. Exams are not only a way to measure what students know, but also a chance for students to reflect on how they studied and what they can do better next time. One way that students can reflect is by using exam wrappers, which are short questions that students answer after they get their graded exams back. A type of AI called natural language processing (NLP) is used in this dissertation, which can analyze text and find patterns and meanings in it. This study also uses a powerful AI tool called GPT-3.5, which can generate text and answer questions. The dissertation has three manuscripts that compare the traditional way of analyzing exam wrappers, which is done by hand, with the new way of using NLP and GPT-3.5, evaluate a specific promising NLP method, and use this method to try and gain a deeper understanding in students self-regulated learning (SRL) while preparing for exams. The data comes from 3,800 exam wrappers from a physics course for engineering students. The first manuscript develops a way of using NLP and GPT-3.5 to find out what learning strategies and goals students talk about in their exam wrappers and compares it to more traditional methods of analysis. The second manuscript tests how accurate a specific NLP technique is in finding these strategies and goals. The third manuscript looks at how different students use different strategies and goals depending on how well they did on the exams using the NLP technique in the second manuscript. I found that NLP and GPT-3.5 can aid in analyzing exam wrappers faster and provide nuanced insights when compared with manual approaches. The dissertation also shows what learning strategies and goals are most discussed for engineering students as they prepare for exams. The dissertation gives some suggestions, challenges, and ideas for future research on AI and learning from exams.

This thesis investigates the impact of finetuning the LLaMA 33B language model on partisan news datasets, revealing negligible changes and underscoring the enduring influence of pretraining datasets on model opinions. Training nine models across nine distinct news datasets spanning three topics and two ideologies, the study found consistent demographic representation, predominantly favoring liberal, college-educated, high-income, and non-religious demographics. Interestingly, a depolarizing effect emerged from partisan news finetuning, suggesting that intense exposure to topic-specific information might lead to depolarization, irrespective of ideological alignment. Despite the exposure to contrasting viewpoints, LLaMA 33B maintained its common sense reasoning ability, showing minimal variance on evaluation metrics like Hellaswag accuracy, ARC accuracy, and TruthfulQA MC1 and MC2. These results might indicate robustness in common sense reasoning or a deficiency in synthesizing diverse contextual information. Ultimately, this thesis demonstrates the resilience of high-performing language models like LLaMA 33B against targeted ideological bias, demonstrating their continued functionality and reasoning ability, even when subjected to highly partisan information environments.

<p dir="ltr">Sexual harassment is regarded as a serious issue in society, with a particularly negative impact on young children and adolescents. Online sexual harassment has recently gained prominence as a significant number of communications have taken place online. Online sexual harassment can happen anywhere in the world because of the global nature of the internet, which transcends geographical barriers and allows people to communicate electronically. Online sexual harassment can occur in a wide variety of environments such as through work mail or chat apps in the workplace, on social media, in online communities, and in games (Chawki & El Shazly, 2013).<br>However, especially for non-native English speakers, due to cultural differences and language barriers, may vary in their understanding or interpretation of text-based sexual harassment (Welsh, Carr, MacQuarrie, & Huntley, 2006). To bridge this gap, previous studies have proposed large language models to detect and classify online sexual harassment, prompting a need to explore how language models comprehend the nuanced aspects of sexual harassment data. Prior to exploring the role of language models, it is critical to recognize the current gaps in knowledge that these models could potentially address in order to comprehend and interpret the complex nature of sexual harassment.</p><p><br></p><p dir="ltr">The Large Language Model (LLM) has attracted significant attention recently due to its exceptional performance on a broad spectrum of tasks. However, these models are characterized by being very sensitive to input data (Fujita et al., 2022; Wei, Wang, et al., 2022). Thus, the purpose of this study is to examine how various LLMs interpret data that falls under the domain of sexual harassment and how they comprehend it after replacing Out-of-Vocabulary words.</p><p dir="ltr"><br>This research examines the impact of Out-of-Vocabulary words on the performance of LLMs in classifying sexual harassment behaviors in text. The study compares the story classification abilities of cutting-edge LLM, before and after the replacement of Out-of-Vocabulary words. Through this investigation, the study provides insights into the flexibility and contextual awareness of LLMs when managing delicate narratives in the context of sexual harassment stories as well as raises awareness of sensitive social issues.</p>

<p dir="ltr">This research explores the application of large language models (LLMs) to comprehension of puns. Leveraging the expansive capabilities of LLMs, this study delves into the domain of pun classification by examining it through the prism of two humor theories: the Computational Model of Humor and the Benign Violation theory, which is an extension of the N+V Theory. The computational model posits that for a phrase to qualify as a pun, it must possess both ambiguity and distinctiveness, characterized by a word that can be interpreted in two plausible ways, each interpretation being supported by at least one unique word. On the other hand, the Benign Violation theory posits that puns work by breaching one linguistic rule while conforming to another, thereby creating a "benign violation." By leveraging the capabilities of large language models (LLMs), this research endeavors to scrutinize a curated collection of English language puns. Our aim is to assess the validity and effectiveness of the use of these theoretical frameworks in accurately classifying puns. We undertake controlled experiments on the dataset, selectively removing a condition specific to one theory and then evaluating the puns based on the criteria of the other theory to see how well it classifies the altered inputs. This approach allows us to uncover deeper insights into the processes that facilitate the recognition of puns and to explore the practical implications of applying humor theories. The findings of our experiments, detailed in the subsequent sections, sheds light on how the alteration of specific conditions impacts the ability of the LLMs to accurately classify puns, according to each theory, where each component of the theory does not influence the result to the same extent, thereby contributing to our understanding of humor mechanics through the eyes of LLMs.</p>

Användningen av Artificiell Intelligens (AI) är utbredd bland verksamma företag idag, likväl privatpersoner. Det har blivit en integrerad del av vårt samhälle som ofta går obemärkt förbi. Allt från face recognition, självkörande bilar och automatisering inom arbetsrelaterade områden, har AI onekligen påverkat omvärlden. I takt med att AI-modeller fortsätter att utvecklas tillkommer även farhågor om dess påverkan på jobb, tillhörande säkerhetsrisker och etiska dilemman. Uppsatsens litteratur hjälper till att skildra AI historiskt, i nutid, men även ge en uppfattning om vart den är på väg. Den AI-modell som i nuläget har väckt störst uppmärksamhet är ChatGPT. Dess potential tycks inte ha några gränser, därmed uppstod relevansen för att öka kunskapen kring AI-modellen. Vidare gjordes en avgränsning, där fokusområdet var att undersöka hur ChatGPT kan generera kodkommentarer och potentiellt agera som ett hjälpmedel vid kommentering av källkod. I samband med avgränsningen och fokusområdet bildades även forskningsfrågan: Large Language Models: Bedömning av ChatGPT:s potential som verktyg för kommentering av kod För att besvara forskningsfrågan har avhandlingen varit baserat på en kvalitativ ansats, där urvalet av respondenter har varit programmerare. Den primära datainsamlingen har genomförts via två semistrukturerade intervjuer, varav den inledande innefattade initiala känslor kring ChatGPT och övergripande fakta om respektive intervjuobjekt. Vidare gjordes det en observation för att få en inblick i hur AI-modellen används av programmerare, för att avslutningsvis göra en uppföljande intervju post-observation i syfte att samla tankarna från intervjuobjekten efter användning av ChatGPT för att generera kodkommentarer. Baserat på den insamlade empirin kunde studien konkludera vissa begränsningar i den nuvarande modellen, inte minst behovet av tydliga instruktioner. Trots brister visar ChatGPTs framställning potential att vara en betydande resurs för kommentering av kod i framtiden. Resultaten indikerar att modellen kan generera relativt passande kommentarer i de analyserade kodkodstycken. Emellertid uttryckte deltagarna under de avslutande intervjuerna generellt sett att kommentarerna var redundanta och saknade betydande värde för att öka förståelsen av källkoden. Respondenterna diskuterade dock möjligheterna att använda ChatGPT i framtiden, men underströk behovet av förbättringar för att göra det till en tillförlitlig metod inom arbetsrelaterade situationer. / The usage of Artificial Intelligence (AI) is widespread among both companies and individuals today. It has become an integrated part of our society, often going unnoticed. From face recognition and self-driving cars to automation in work-related areas, AI has undeniably impacted the world. As AI models continue to evolve, concerns about their impact on jobs, associated security risks, and ethical dilemmas arise. The literature in this essay helps portray AI historically, in the present, and provides an insight into its future direction. The AI model that has currently garnered the most attention is ChatGPT. Its potential seems limitless, which prompted the relevance of increasing knowledge about the AI model. Furthermore, a delimitation was made, where the focus area was to investigate how ChatGPT can generate code comments and potentially act as a tool for commenting source code. As part of the research focus and scope, the research question was formulated: "Large Language Models: Assessment of ChatGPT's Potential as a Tool for Code Commenting." To answer the research question, the thesis adopted a qualitative approach, with programmers as the selected respondents. The primary data collection was conducted through two semi-structured interviews, where the initial interview involved capturing initial impressions of ChatGPT and gathering general information about the interviewees. Additionally, an observation was carried out to gain insights into how programmers utilize the AI model, followed by a post-observation interview to gather the interviewees' thoughts after using ChatGPT to generate code comments. Based on the collected empirical data, the study was able to conclude certain limitations in the current model, particularly the need for clear instructions. Despite these limitations, ChatGPT's performance demonstrates the potential to be a significant resource for code commenting in the future. The results indicate that the model can generate relatively suitable comments in the analyzed code snippets. However, during the concluding interviews, participants generally expressed that the comments were redundant and lacked significant value in enhancing the understanding of the source code. Nevertheless, the respondents 2 discussed the possibilities of using ChatGPT in the future, while emphasizing the need for improvements to establish it as a reliable method in work-related situations.

The knowledge of tests conducted on a problematic vehicle is essential for enhancing the efficiency of mechanics. Therefore, identifying the tests performed in each repair case is of utmost importance. This thesis explores techniques for extracting data from unstructured repair reports to identify component tests. The main emphasis is on developing a supervised multi-class classifier to categorize data and extract sentences that describe repair diagnoses and actions. It has been shown that incorporating a category-aware contrastive learning objective can improve the repair report classifier’s performance. The proposed approach involves training a sentence representation model based on a pre-trained model using a category-aware contrastive learning objective. Subsequently, the sentence representation model is further trained on the classification task using a loss function that combines the cross-entropy and supervised contrastive learning losses. By applying this method, the macro F1-score on the test set is increased from 90.45 to 90.73. The attempt to enhance the performance of the repair report classifier using a noisy data classifier proves unsuccessful. The noisy data classifier is trained using a prompt-based fine-tuning method, incorporating open-ended questions and two examples in the prompt. This approach achieves an F1-score of 91.09 and the resulting repair report classification datasets are found easier to classify. However, they do not contribute to an improvement in the repair report classifier’s performance. Ultimately, the repair report classifier is utilized to aid in creating the input necessary for identifying component tests. An information retrieval method is used to conduct the test identification. The incorporation of this classifier and the existing labels when creating queries leads to an improvement in the mean average precision at the top 3, 5, and 10 positions by 0.62, 0.81, and 0.35, respectively, although with a slight decrease of 0.14 at the top 1 position.

As advancements in chatbots and Large Language Models (LLMs) such as GPT-3.5 and GPT-4 continue, their applications in diverse fields, including psychology, expand. This study investigates the effectiveness of LLMs optimized through prompt engineering, aiming to enhance their performance in psychological applications. To this end, two distinct versions of a GPT-3.5-based chatbot were developed: a version similar to the base model, and a version equipped with a more extensive system prompt detailing expected behavior. A panel of professional psychologists evaluated these models based on a predetermined set of questions, providing insight into their potential future use as psychological tools. Our results indicate that an overly prescriptive system prompt can unintentionally limit the versatility of the chatbot, making a careful balance in instruction specificity essential. Furthermore, while our study suggests that current LLMs such as GPT-3.5 are not capable of fully replacing human psychologists, they can provide valuable assistance in tasks such as basic question answering, consolation and validation, and triage. These findings provide a foundation for future research into the effective integration of LLMs in psychology and contribute valuable insights into the promising field of AI-assisted psychological services. / I takt med att framstegen inom chatbots och stora språkmodeller (LLMs) som GPT-3.5 och GPT-4 fortsätter utvidgas deras potentiella tillämpningar inom olika områden, inklusive psykologi. Denna studie undersöker effektiviteten av LLMs optimerade genom prompt engineering, med målet att förbättra deras prestanda inom psykologiska tillämpningar. I detta syfte utvecklades två distinkta versioner av en chatbot baserad på GPT-3.5: en version som liknar bas-modellen, och en version utrustad med en mer omfattande systemprompt som detaljerar förväntat beteende. En panel av professionella psykologer utvärderade dessa modeller baserat på en förbestämd uppsättning frågor, vilket ger inblick i deras potentiella framtida användning som psykologiska verktyg. Våra resultat tyder på att en överdrivet beskrivande systemprompt kan ofrivilligt begränsa chatbotens mångsidighet, vilket kräver en noggrann balans i specificiteten av prompten. Vidare antyder vår studie att nuvarande LLMs som GPT-3.5 inte kan ersätta mänskliga psykologer helt och hållet, men att de kan ge värdefull hjälp i uppgifter som grundläggande frågebesvaring, tröst och bekräftelse, samt triage. Dessa resultat ger en grund för framtida forskning om effektiv integration av LLMs inom psykologi och bidrar med värdefulla insikter till det lovande fältet av AI-assisterade psykologtjänster.