Low technology and eHealth literacy among older adults in retirement communities hinder engagement with digital tools. To address this, we designed an LLM-powered chatbot prototype using a human-centered approach for a local retirement community. Through interviews and persona development, we prioritized accessibility and dual functionality: simplifying internal information retrieval and improving technology and eHealth literacy. A pilot trial with residents demonstrated high satisfaction and ease of use, but also identified areas for further improvement. Based on the feedback, we refined the chatbot using GPT-3.5 Turbo and Streamlit. The chatbot employs tailored prompt engineering to deliver concise responses. Accessible features like adjustable font size, interface theme and personalized follow-up responses were implemented. Future steps include enabling voice-to-text function and longitudinal intervention studies. Together, our results highlight the potential of LLM-driven chatbots to empower older adults through accessible, personalized interactions, bridging literacy gaps in retirement communities.
Multimodal Large Language Models (MLLMs) are set to transform how machines process and generate human-like responses by integrating diverse modalities such as text, images, and code. Yet, effectively harnessing their capabilities hinges on optimal prompt engineering. We present a comprehensive experimental evaluation of seven prompt engineering methods applied to 13 open-source MLLMs over 24 tasks spanning Reasoning and Compositionality, Multimodal Understanding and Alignment, Complex Code Generation and Execution, and Knowledge Retrieval and Integration. Our approach stratifies models by parameter count into Small (<4B), Medium (4B-10B), and Large (>10B) categories and compares prompting techniques including Zero-Shot, One-Shot, Few-Shot, Chain-of-Thought, Analogical, Generated Knowledge, and Tree-of-Thought. While Large MLLMs excel in structured tasks such as code generation, achieving accuracies up to 96.88% under Few-Shot prompting, all models struggle with complex reasoning and abstract understanding, often yielding accuracies below 60% and high hallucination rates. Structured reasoning prompts frequently increased hallucination up to 75% in small models and led to longer response times (over 20 seconds in Large MLLMs), while simpler prompting methods provided more concise and efficient outputs. No single prompting method uniformly optimises all task types. Instead, adaptive strategies combining example-based guidance with selective structured reasoning are essential to enhance robustness, efficiency, and factual accuracy. Our findings offer practical recommendations for prompt engineering and support more reliable deployment of MLLMs across applications including AI-assisted coding, knowledge retrieval, and multimodal content understanding.
Generative agents have been increasingly used to simulate human behaviour in silico, driven by large language models (LLMs). These simulacra serve as sandboxes for studying human behaviour without compromising privacy or safety. However, it remains unclear whether such agents can truly represent real individuals. This work compares survey data from the Understanding America Study (UAS) on healthcare decision-making with simulated responses from generative agents. Using demographic-based prompt engineering, we create digital twins of survey respondents and analyse how well different LLMs reproduce real-world behaviours. Our findings show that some LLMs fail to reflect realistic decision-making, such as predicting universal vaccine acceptance. However, Llama 3 captures variations across race and Income more accurately but also introduces biases not present in the UAS data. This study highlights the potential of generative agents for behavioural research while underscoring the risks of bias from both LLMs and prompting strategies.
Large language models (LLMs) have shown potential in supporting decision-making applications, particularly as personal conversational assistants in the financial, healthcare, and legal domains. While prompt engineering strategies have enhanced the capabilities of LLMs in decision-making, cognitive biases inherent to LLMs present significant challenges. Cognitive biases are systematic patterns of deviation from norms or rationality in decision-making that can lead to the production of inaccurate outputs. Existing cognitive bias mitigation strategies assume that input prompts contain (exactly) one type of cognitive bias and therefore fail to perform well in realistic settings where there maybe any number of biases. To fill this gap, we propose a cognitive debiasing approach, called self-debiasing, that enhances the reliability of LLMs by iteratively refining prompts. Our method follows three sequential steps -- bias determination, bias analysis, and cognitive debiasing -- to iteratively mitigate potential cognitive biases in prompts. Experimental results on finance, healthcare, and legal decision-making tasks, using both closed-source and open-source LLMs, demonstrate that the proposed self-debiasing method outperforms both advanced prompt engineering methods and existing cognitive debiasing techniques in average accuracy under no-bias, single-bias, and multi-bias settings.
Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, with their performance heavily dependent on the quality of input prompts. While prompt engineering has proven effective, it typically relies on manual adjustments, making it time-consuming and potentially suboptimal. This paper introduces GAAPO (Genetic Algorithm Applied to Prompt Optimization), a novel hybrid optimization framework that leverages genetic algorithm principles to evolve prompts through successive generations. Unlike traditional genetic approaches that rely solely on mutation and crossover operations, GAAPO integrates multiple specialized prompt generation strategies within its evolutionary framework. Through extensive experimentation on diverse datasets including ETHOS, MMLU-Pro, and GPQA, our analysis reveals several important point for the future development of automatic prompt optimization methods: importance of the tradeoff between the population size and the number of generations, effect of selection methods on stability results, capacity of different LLMs and especially reasoning models to be able to automatically generate prompts from similar queries... Furthermore, we provide insights into the relative effectiveness of different prompt generation strategies and their evolution across optimization phases. These findings contribute to both the theoretical understanding of prompt optimization and practical applications in improving LLM performance.
Large Language Models (LLMs) have demonstrated remarkable capabilities in code generation. However, the quality of the generated code is heavily dependent on the structure and composition of the prompts used. Crafting high-quality prompts is a challenging task that requires significant knowledge and skills of prompt engineering. To advance the automation support for the prompt engineering for LLM-based code generation, we propose a novel solution Diffusion-Driven Prompt Tuning (DDPT) that learns how to generate optimal prompt embedding from Gaussian Noise to automate the prompt engineering for code generation. We evaluate the feasibility of diffusion-based optimization and abstract the optimal prompt embedding as a directional vector toward the optimal embedding. We use the code generation loss given by the LLMs to help the diffusion model capture the distribution of optimal prompt embedding during training. The trained diffusion model can build a path from the noise distribution to the optimal distribution at the sampling phrase, the evaluation result demonstrates that DDPT helps improve the prompt optimization for code generation.
The integration of Large Language Models (LLMs) with Learning Management Systems (LMSs) has the potential to enhance task automation and accessibility in education. However, hallucination where LLMs generate inaccurate or misleading information remains a significant challenge. This study introduces the Dynamic Course Content Integration (DCCI) mechanism, which dynamically retrieves and integrates course content and curriculum from Canvas LMS into the LLM-powered assistant, Ask ME. By employing prompt engineering to structure retrieved content within the LLM's context window, DCCI ensures accuracy, relevance, and contextual alignment, mitigating hallucination. To evaluate DCCI's effectiveness, Ask ME's usability, and broader student perceptions of AI in education, a mixed-methods approach was employed, incorporating user satisfaction ratings and a structured survey. Results from a pilot study indicate high user satisfaction (4.614/5), with students recognizing Ask ME's ability to provide timely and contextually relevant responses for both administrative and course-related inquiries. Additionally, a majority of students agreed that Ask ME's integration with course content in Canvas LMS reduced platform-switching, improving usability, engagement, and comprehension. AI's role in reducing classroom hesitation and fostering self-directed learning and intellectual curiosity was also highlighted. Despite these benefits and positive perception of AI tools, concerns emerged regarding over-reliance on AI, accuracy limitations, and ethical issues such as plagiarism and reduced student-teacher interaction. These findings emphasize the need for strategic AI implementation, ethical safeguards, and a pedagogical framework that prioritizes human-AI collaboration over substitution.
Understanding and mitigating biases is critical for the adoption of large language models (LLMs) in high-stakes decision-making. We introduce Admissions and Hiring, decision tasks with hypothetical applicant profiles where a person's race can be inferred from their name, as simplified test beds for racial bias. We show that Gemma 2B Instruct and LLaMA 3.2 3B Instruct exhibit strong biases. Gemma grants admission to 26% more White than Black applicants, and LLaMA hires 60% more Asian than White applicants. We demonstrate that these biases are resistant to prompt engineering: multiple prompting strategies all fail to promote fairness. In contrast, using distributed alignment search, we can identify "race subspaces" within model activations and intervene on them to debias model decisions. Averaging the representation across all races within the subspaces reduces Gemma's bias by 37-57%. Finally, we examine the generalizability of Gemma's race subspaces, and find limited evidence for generalization, where changing the prompt format can affect the race representation. Our work suggests mechanistic approaches may provide a promising venue for improving the fairness of LLMs, but a universal race representation remains elusive.
Large Reasoning Models (LRMs) such as DeepSeek-R1 and OpenAI o1 have demonstrated remarkable capabilities in various reasoning tasks. Their strong capability to generate and reason over intermediate thoughts has also led to arguments that they may no longer require extensive prompt engineering or optimization to interpret human instructions and produce accurate outputs. In this work, we aim to systematically study this open question, using the structured task of event extraction for a case study. We experimented with two LRMs (DeepSeek-R1 and o1) and two general-purpose Large Language Models (LLMs) (GPT-4o and GPT-4.5), when they were used as task models or prompt optimizers. Our results show that on tasks as complicated as event extraction, LRMs as task models still benefit from prompt optimization, and that using LRMs as prompt optimizers yields more effective prompts. Finally, we provide an error analysis of common errors made by LRMs and highlight the stability and consistency of LRMs in refining task instructions and event guidelines.
With the increasingly deep integration of large language models (LLMs) across diverse domains, the effectiveness of their safety mechanisms is encountering severe challenges. Currently, jailbreak attacks based on prompt engineering have become a major safety threat. However, existing methods primarily rely on black-box manipulation of prompt templates, resulting in poor interpretability and limited generalization. To break through the bottleneck, this study first introduces the concept of Defense Threshold Decay (DTD), revealing the potential safety impact caused by LLMs' benign generation: as benign content generation in LLMs increases, the model's focus on input instructions progressively diminishes. Building on this insight, we propose the Sugar-Coated Poison (SCP) attack paradigm, which uses a "semantic reversal" strategy to craft benign inputs that are opposite in meaning to malicious intent. This strategy induces the models to generate extensive benign content, thereby enabling adversarial reasoning to bypass safety mechanisms. Experiments show that SCP outperforms existing baselines. Remarkably, it achieves an average attack success rate of 87.23% across six LLMs. For defense, we propose Part-of-Speech Defense (POSD), leveraging verb-noun dependencies for syntactic analysis to enhance safety of LLMs while preserving their generalization ability.
3D object detection is crucial for autonomous driving, leveraging both LiDAR point clouds for precise depth information and camera images for rich semantic information. Therefore, the multi-modal methods that combine both modalities offer more robust detection results. However, efficiently fusing LiDAR points and images remains challenging due to the domain gaps. In addition, the performance of many models is limited by the amount of high quality labeled data, which is expensive to create. The recent advances in foundation models, which use large-scale pre-training on different modalities, enable better multi-modal fusion. Combining the prompt engineering techniques for efficient training, we propose the Prompted Foundational 3D Detector (PF3Det), which integrates foundation model encoders and soft prompts to enhance LiDAR-camera feature fusion. PF3Det achieves the state-of-the-art results under limited training data, improving NDS by 1.19% and mAP by 2.42% on the nuScenes dataset, demonstrating its efficiency in 3D detection.
Large Language Models (LLMs) equipped with web search capabilities have demonstrated impressive potential for deep research tasks. However, current approaches predominantly rely on either manually engineered prompts (prompt engineering-based) with brittle performance or reinforcement learning within controlled Retrieval-Augmented Generation (RAG) environments (RAG-based) that fail to capture the complexities of real-world interaction. In this paper, we introduce DeepResearcher, the first comprehensive framework for end-to-end training of LLM-based deep research agents through scaling reinforcement learning (RL) in real-world environments with authentic web search interactions. Unlike RAG-based approaches that assume all necessary information exists within a fixed corpus, our method trains agents to navigate the noisy, unstructured, and dynamic nature of the open web. We implement a specialized multi-agent architecture where browsing agents extract relevant information from various webpage structures and overcoming significant technical challenges. Extensive experiments on open-domain research tasks demonstrate that DeepResearcher achieves substantial improvements of up to 28.9 points over prompt engineering-based baselines and up to 7.2 points over RAG-based RL agents. Our qualitative analysis reveals emergent cognitive behaviors from end-to-end RL training, including the ability to formulate plans, cross-validate information from multiple sources, engage in self-reflection to redirect research, and maintain honesty when unable to find definitive answers. Our results highlight that end-to-end training in real-world web environments is not merely an implementation detail but a fundamental requirement for developing robust research capabilities aligned with real-world applications. We release DeepResearcher at https://github.com/GAIR-NLP/DeepResearcher.
Zero-shot text classification typically relies on prompt engineering, but the inherent prompt brittleness of large language models undermines its reliability. Minor changes in prompt can cause significant discrepancies in model performance. We attribute this prompt brittleness largely to the narrow focus on nexttoken probabilities in existing methods. To address this, we propose Placeholding Parallel Prediction (P3), a novel approach that predicts token probabilities across multiple positions and simulates comprehensive sampling of generation paths in a single run of a language model. Experiments show improved accuracy and up to 98% reduction in the standard deviation across prompts, boosting robustness. Even without a prompt, P3 maintains comparable performance, reducing the need for prompt engineering.
Natural Language to SQL (NL2SQL) has emerged as a critical task for enabling seamless interaction with databases. Recent advancements in Large Language Models (LLMs) have demonstrated remarkable performance in this domain. However, existing NL2SQL methods predominantly rely on closed-source LLMs leveraging prompt engineering, while open-source models typically require fine-tuning to acquire domain-specific knowledge. Despite these efforts, open-source LLMs struggle with complex NL2SQL tasks due to the indirect expression of user query objectives and the semantic gap between user queries and database schemas. Inspired by the application of reinforcement learning in mathematical problem-solving to encourage step-by-step reasoning in LLMs, we propose LearNAT (Learning NL2SQL with AST-guided Task Decomposition), a novel framework that improves the performance of open-source LLMs on complex NL2SQL tasks through task decomposition and reinforcement learning. LearNAT introduces three key components: (1) a Decomposition Synthesis Procedure that leverages Abstract Syntax Trees (ASTs) to guide efficient search and pruning strategies for task decomposition, (2) Margin-aware Reinforcement Learning, which employs fine-grained step-level optimization via DPO with AST margins, and (3) Adaptive Demonstration Reasoning, a mechanism for dynamically selecting relevant examples to enhance decomposition capabilities. Extensive experiments on two benchmark datasets, Spider and BIRD, demonstrate that LearNAT enables a 7B-parameter open-source LLM to achieve performance comparable to GPT-4, while offering improved efficiency and accessibility.
Large language models (LLMs) have created new opportunities to assist teachers and support student learning. Methods such as chain-of-thought (CoT) prompting enable LLMs to grade formative assessments in science, providing scores and relevant feedback to students. However, the extent to which these methods generalize across curricula in multiple domains (such as science, computing, and engineering) remains largely untested. In this paper, we introduce Chain-of-Thought Prompting + Active Learning (CoTAL), an LLM-based approach to formative assessment scoring that (1) leverages Evidence-Centered Design (ECD) principles to develop curriculum-aligned formative assessments and rubrics, (2) applies human-in-the-loop prompt engineering to automate response scoring, and (3) incorporates teacher and student feedback to iteratively refine assessment questions, grading rubrics, and LLM prompts for automated grading. Our findings demonstrate that CoTAL improves GPT-4's scoring performance, achieving gains of up to 24.5% over a non-prompt-engineered baseline. Both teachers and students view CoTAL as effective in scoring and explaining student responses, each providing valuable refinements to enhance grading accuracy and explanation quality.
Recent advances in large language and vision-language models have enabled zero-shot inference, allowing models to solve new tasks without task-specific training. Various adaptation techniques such as prompt engineering, In-Context Learning (ICL), and supervised fine-tuning can further enhance the model's performance on a downstream task, but they require substantial manual effort to construct effective prompts or labeled examples. In this work, we introduce a joint inference framework for fully unsupervised adaptation, eliminating the need for manual prompt engineering and labeled examples. Unlike zero-shot inference, which makes independent predictions, the joint inference makes predictions simultaneously for all inputs in a given task. Since direct joint inference involves computationally expensive optimization, we develop efficient approximation techniques, leading to two unsupervised adaptation methods: unsupervised fine-tuning and unsupervised ICL. We demonstrate the effectiveness of our methods across diverse tasks and models, including language-only Llama-3.1 on natural language processing tasks, reasoning-oriented Qwen2.5-Math on grade school math problems, vision-language OpenFlamingo on vision tasks, and the API-only access GPT-4o model on massive multi-discipline tasks. Our experiments demonstrate substantial improvements over the standard zero-shot approach, including 39% absolute improvement on the challenging GSM8K math reasoning dataset. Remarkably, despite being fully unsupervised, our framework often performs on par with supervised approaches that rely on ground truth labels.
Large Language Models (LLMs) have significantly advanced Natural Language Processing (NLP), particularly in Natural Language Understanding (NLU) tasks. As we progress toward an agentic world where LLM-based agents autonomously handle specialized tasks, it becomes crucial for these models to adapt to new tasks without forgetting previously learned information - a challenge known as catastrophic forgetting. This study evaluates the continual fine-tuning of various open-source LLMs with different parameter sizes (specifically models under 10 billion parameters) on key NLU tasks from the GLUE benchmark, including SST-2, MRPC, CoLA, and MNLI. By employing prompt engineering and task-specific adjustments, we assess and compare the models' abilities to retain prior knowledge while learning new tasks. Our results indicate that models such as Phi-3.5-mini exhibit minimal forgetting while maintaining strong learning capabilities, making them well-suited for continual learning environments. Additionally, models like Orca-2-7b and Qwen2.5-7B demonstrate impressive learning abilities and overall performance after fine-tuning. This work contributes to understanding catastrophic forgetting in LLMs and highlights prompting engineering to optimize model performance for continual learning scenarios.
Synthetic data offers a promising path to train models while preserving data privacy. Differentially private (DP) finetuning of large language models (LLMs) as data generator is effective, but is impractical when computation resources are limited. Meanwhile, prompt-based methods such as private evolution, depend heavily on the manual prompts, and ineffectively use private information in their iterative data selection process. To overcome these limitations, we propose CTCL (Data Synthesis with ConTrollability and CLustering), a novel framework for generating privacy-preserving synthetic data without extensive prompt engineering or billion-scale LLM finetuning. CTCL pretrains a lightweight 140M conditional generator and a clustering-based topic model on large-scale public data. To further adapt to the private domain, the generator is DP finetuned on private data for fine-grained textual information, while the topic model extracts a DP histogram representing distributional information. The DP generator then samples according to the DP histogram to synthesize a desired number of data examples. Evaluation across five diverse domains demonstrates the effectiveness of our framework, particularly in the strong privacy regime. Systematic ablation validates the design of each framework component and highlights the scalability of our approach.
Vision-language models (VLMs) have advanced rapidly in processing multimodal information, but their ability to reconcile conflicting signals across modalities remains underexplored. This work investigates how VLMs process ASCII art, a unique medium where textual elements collectively form visual patterns, potentially creating semantic-visual conflicts. We introduce a novel evaluation framework that systematically challenges five state-of-the-art models (including GPT-4o, Claude, and Gemini) using adversarial ASCII art, where character-level semantics deliberately contradict global visual patterns. Our experiments reveal a strong text-priority bias: VLMs consistently prioritize textual information over visual patterns, with visual recognition ability declining dramatically as semantic complexity increases. Various mitigation attempts through visual parameter tuning and prompt engineering yielded only modest improvements, suggesting that this limitation requires architectural-level solutions. These findings uncover fundamental flaws in how current VLMs integrate multimodal information, providing important guidance for future model development while highlighting significant implications for content moderation systems vulnerable to adversarial examples.
Geographic Question Answering (GeoQA) addresses natural language queries in geographic domains to fulfill complex user demands and improve information retrieval efficiency. Traditional QA systems, however, suffer from limited comprehension, low retrieval accuracy, weak interactivity, and inadequate handling of complex tasks, hindering precise information acquisition. This study presents GeoRAG, a knowledge-enhanced QA framework integrating domain-specific fine-tuning and prompt engineering with Retrieval-Augmented Generation (RAG) technology to enhance geographic knowledge retrieval accuracy and user interaction. The methodology involves four components: (1) A structured geographic knowledge base constructed from 3267 corpora (research papers, monographs, and technical reports), categorized via a multi-agent approach into seven dimensions: semantic understanding, spatial location, geometric morphology, attribute characteristics, feature relationships, evolutionary processes, and operational mechanisms. This yielded 145234 classified entries and 875432 multi-dimensional QA pairs. (2) A multi-label text classifier based on BERT-Base-Chinese, trained to analyze query types through geographic dimension classification. (3) A retrieval evaluator leveraging QA pair data to assess query-document relevance, optimizing retrieval precision. (4) GeoPrompt templates engineered to dynamically integrate user queries with retrieved information, enhancing response quality through dimension-specific prompting. Comparative experiments demonstrate GeoRAG's superior performance over conventional RAG across multiple base models, validating its generalizability. This work advances geographic AI by proposing a novel paradigm for deploying large language models in domain-specific contexts, with implications for improving GeoQA systems scalability and accuracy in real-world applications.