Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 31 for “"Prompt design"”.

  1. Investigating Zero-Shot Diagnostic Pathology in Vision-Language Models with Efficient Prompt Design

    … to the linguistic structure and specificity of prompts, which can significantly impact diagnostic accuracy, reproducibility, and clinical interpretability. This thesis systematically investigates the role of prompt engineering in enhancing the diagnostic performance of VLMs in histopathology. We …

    concordia Download PDF for Investigating Zero-Shot Diagnostic Pathology in Vision-Language Models with Efficient Prompt Design (opens in a new tab)

  2. Evaluating LLMs for CPE Identification in IoT Reconnaissance

    … performance varies with model scale, family, and prompt design. I evaluate 36 LLMs across three deployment tiers (frontier hosted APIs, Ollama Cloud, and local vLLM), plus Nmap’s native CPE extraction as the no-LLM baseline, on a curated corpus of Nmap covering 16 analyzed IoT device entries drawn …

    eastern-wash Repository record for Evaluating LLMs for CPE Identification in IoT Reconnaissance (opens in a new tab)

  3. Automating Financial Report Analysis and Generation using LLMs

    … Still, their application and research on designing them in the financial domain remains limited. Objective: This study explored how LLMs can be integrated into financial report analysis and commentary generation, with a particular focus on the software engineering challenges encountered …

    chalmers Repository record for Automating Financial Report Analysis and Generation using LLMs (opens in a new tab)

  4. AI Feedback vs. Teacher Feedback: A Comparative Study of English-Speaking Practice among L1 Mandarin EFL Learners

    … IELTS Speaking programme. Using a mixed methods design, the study involved 24 participants, with 13 receiving AI-supported feedback through Doubao-Seed-2.0, a generative AI chatbot, and 11 receiving teacher-led feedback. Outcomes were examined through official IELTS Speaking scores and learner …

    auckland-tech Repository record for AI Feedback vs. Teacher Feedback: A Comparative Study of English-Speaking Practice among L1 Mandarin EFL Learners (opens in a new tab)

  5. Semantics-Behavior Coupled Bayesian Optimization for Efficient Black-Box Prompt Search

    … language models (LLMs) are highly sensitive to prompt design, making prompt optimization important in black-box settings where models are accessible only through API calls. This thesis proposes DualBO, a semantics-behavior coupled Bayesian optimization framework for efficient prompt search. …

    waterloo Repository record for Semantics-Behavior Coupled Bayesian Optimization for Efficient Black-Box Prompt Search (opens in a new tab)

  6. Automatic Measurement of Dialogue Engagingness in Multilingual Settings

    … history, are complicated to implement, or are designed for a specific dataset. Moreover, they have been tested exclusively on English conversations. Given that dialogue systems are increasingly available in languages beyond English, it is important to evaluate systems in more than one language. …

    portland-state Repository record for Automatic Measurement of Dialogue Engagingness in Multilingual Settings (opens in a new tab)

  7. Think or Step-by-Step? UnZIPping the Black Box in Zero-Shot Prompts

    Zero-shot instructional prompts like “Let’s think step-by-step” have dramatically improved Large Language Model (LLM) performance, yet we lack systematic understanding of which words in these prompts drive their effectiveness. Is think or step-by-step more crucial? How much do seemingly minor word …

    brock Repository record for Think or Step-by-Step? UnZIPping the Black Box in Zero-Shot Prompts (opens in a new tab)

  8. Evaluating Function Calling in Language Models through a Conversational Agent

    … of GPT-4o and GPT-4o-mini is conducted across prompt specificity, query difficulty, and query type. Results show that function-calling accuracy improves significantly with more explicit and structured prompts, with GPT-4o achieving the highest performance. Spatial and constraint-based queries …

    vt Repository record for Evaluating Function Calling in Language Models through a Conversational Agent (opens in a new tab)

  9. Adaptive and Generalizable Vision-Language Models

    … inference. In this work, we introduce a domain prompt fusion framework aimed at improving the generalization capability of CLIP-based models under domain shift. Our approach integrates three core components: a dual-part soft prompt (comprising domain-agnostic and domain-specific prompts), a …

    chalmers Repository record for Adaptive and Generalizable Vision-Language Models (opens in a new tab)

  10. From invoices to insights : feature engineering strategies for multilingual product classification in medical device procurement

    … subsidiaries. For the LLM, retrieval-augmented prompting lifts accuracy from 53% to 91%, while prompt design contributes only ~3 points. The two approaches are complementary — ML offers calibrated confidence (ECE ≤ 0.038), while the LLM provides catalogue-wide coverage including unseen item …

    reykjavik Repository record for From invoices to insights : feature engineering strategies for multilingual product classification in medical device procurement (opens in a new tab)

  11. Bridging the Data Gap: LLM-Driven Synthetic Data Generation for Low-Resource Natural Language Inference

    … shaped by factors including model architecture, prompt design, and the inherent complexity of the task itself. These findings show that while LLM-generated synthetic data can mitigate data scarcity issues, annotation quality remains a critical factor influencing model performance. The thesis …

    maynooth Download PDF for Bridging the Data Gap: LLM-Driven Synthetic Data Generation for Low-Resource Natural Language Inference (opens in a new tab)

  12. Analysis of Method Name Prediction in the Era of Large Language Model

    … of large language models (LLMs) introduces a prompt-based alternative where inference can be performed without task-specific retraining. This thesis studies method name prediction in the LLM era by reviewing prior supervised paradigms and by evaluating prompt-based methods through a systematic …

    unsw Repository record for Analysis of Method Name Prediction in the Era of Large Language Model (opens in a new tab)

  13. Using Natural Language Processing to Support Software Security Requirement Development

    … As part of the experiments, hypothesis and prompt design were also examined for their effects upon the performance of these models. Three datasets were used in this study, one composed of security related software requirements, one composed of non-security related requirements, and one …

    rit Repository record for Using Natural Language Processing to Support Software Security Requirement Development (opens in a new tab)

  14. Instruments of Belief: The Victory Box

    … an AI system I call Lucid, which contributed to prompt design, theoretical discussion, and the broader shaping of the work’s fictional framework. Rather than addressing these issues didactically, the installation stages them as an embodied encounter. It records viewers, transforms their images, …

    waterloo Repository record for Instruments of Belief: The Victory Box (opens in a new tab)

  15. A HYBRID RULE-BASED AND LARGE LANGUAGE MODEL FRAMEWORK FOR EXTRACTING ACRONYM–DEFINITION PAIRS FROM SCIENTIFIC LITERATURE

    … Claude Haiku 4.5, and Kimi K2.5) under identical prompts, so differences reflect model capability rather than prompt design. Each LLM provider is evaluated on three input types (the raw PDF, unprocessed extracted text, and preprocessed text) and on seven method configurations spanning the parser …

    rowan Repository record for A HYBRID RULE-BASED AND LARGE LANGUAGE MODEL FRAMEWORK FOR EXTRACTING ACRONYM–DEFINITION PAIRS FROM SCIENTIFIC LITERATURE (opens in a new tab)

  16. Software to Generate Linux Executables for Sending Eiffel Events Based on JSON Schemas at Nexer - Schema-driven code generation using large language models

    … combinations involved. This thesis presents the design and implementation of a software system that auto mates the generation of Linux-executable command-line tools for creating, validating, and sending Eiffel events via RabbitMQ. The system uses a schema-driven approach where the official Eiffel …

    chalmers Repository record for Software to Generate Linux Executables for Sending Eiffel Events Based on JSON Schemas at Nexer - Schema-driven code generation using large language models (opens in a new tab)

  17. Αλγόριθμοι και τεχνικές συγχρονισμού ασύρματων ψηφιακών δεκτών ορθογώνιας γραμμικής διαμόρφωσης

    … addresses critical algorithmic and architectural design issues concerning digital receivers for high-capacity radio communication systems. The results presented herein, in effect, provide a comprehensive approach in optimal pulse shaping networks, channel estimation and synchronization techniques …

    greece Repository record for Αλγόριθμοι και τεχνικές συγχρονισμού ασύρματων ψηφιακών δεκτών ορθογώνιας γραμμικής διαμόρφωσης (opens in a new tab)

  18. Epidemic Modeling with Generative Agents: Methodology, Prompt Sensitivity, and LLM Sensitivity

    … methodology, characterizes its sensitivity to prompt design, and characterizes its sensitivity to the underlying LLM. Three studies — conducted within a shared epidemic GABM framework in which generative agents decide each day whether to stay home or go out — support the central claim: GABM is …

    vt Repository record for Epidemic Modeling with Generative Agents: Methodology, Prompt Sensitivity, and LLM Sensitivity (opens in a new tab)

  19. Prompting discussion : writing prompts, habits of mind, and the shape of the writing classroom.

    … writing classroom and what role writing prompts play in ameliorating or exacerbating those difficulties. It does so by tracing across four disciplines (philosophy, psychology, nursing, and biology) several common conflicts and double binds encountered at the intersection between ideal …

    louisville Repository record for Prompting discussion : writing prompts, habits of mind, and the shape of the writing classroom. (opens in a new tab)

  20. Designing Human- and AI-Assisted Health Coaching for Preventive Behaviour Change

    … how health coaching interventions can be designed, delivered, and evaluated to support preventive health behaviour change through six studies spanning MI and AI-assisted health coaching. Methods: Six studies were conducted: a systematic review and meta-analysis, a pilot randomized …

    mcmaster Repository record for Designing Human- and AI-Assisted Health Coaching for Preventive Behaviour Change (opens in a new tab)

Page 1 of 2