The proliferation of large language models (LLMs) requires robust evaluation of their alignment with local values and ethical standards, especially as existing benchmarks often reflect the cultural, legal, and ideological values of their creators. \textsc{LocalValueBench}, introduced in this paper, is an extensible benchmark designed to assess LLMs' adherence to Australian values, and provides a framework for regulators worldwide to develop their own LLM benchmarks for local value alignment. Employing a novel typology for ethical reasoning and an interrogation approach, we curated comprehensive questions and utilized prompt engineering strategies to probe LLMs' value alignment. Our evaluation criteria quantified deviations from local values, ensuring a rigorous assessment process. Comparative analysis of three commercial LLMs by USA vendors revealed significant insights into their effectiveness and limitations, demonstrating the critical importance of value alignment. This study offers valuable tools and methodologies for regulators to create tailored benchmarks, highlighting avenues for future research to enhance ethical AI development.
The emergent abilities of large language models (LLMs) have demonstrated great potential in solving medical questions. They can possess considerable medical knowledge, but may still hallucinate and are inflexible in the knowledge updates. While Retrieval-Augmented Generation (RAG) has been proposed to enhance the medical question-answering capabilities of LLMs with external knowledge bases, it may still fail in complex cases where multiple rounds of information-seeking are required. To address such an issue, we propose iterative RAG for medicine (i-MedRAG), where LLMs can iteratively ask follow-up queries based on previous information-seeking attempts. In each iteration of i-MedRAG, the follow-up queries will be answered by a conventional RAG system and they will be further used to guide the query generation in the next iteration. Our experiments show the improved performance of various LLMs brought by i-MedRAG compared with conventional RAG on complex questions from clinical vignettes in the United States Medical Licensing Examination (USMLE), as well as various knowledge tests in the Massive Multitask Language Understanding (MMLU) dataset. Notably, our zero-shot i-MedRAG outperforms all existing prompt engineering and fine-tuning methods on GPT-3.5, achieving an accuracy of 69.68% on the MedQA dataset. In addition, we characterize the scaling properties of i-MedRAG with different iterations of follow-up queries and different numbers of queries per iteration. Our case studies show that i-MedRAG can flexibly ask follow-up queries to form reasoning chains, providing an in-depth analysis of medical questions. To the best of our knowledge, this is the first-of-its-kind study on incorporating follow-up queries into medical RAG. The implementation of i-MedRAG is available at https://github.com/Teddy-XiongGZ/MedRAG.
Foundation models have had a big impact in recent years and billions of dollars are being invested in them in the current AI boom. The more popular ones, such as Chat-GPT, are trained on large amounts of data from the Internet, and then reinforcement learning, RAG, prompt engineering and cognitive modelling are used to fine-tune and augment their behavior. This technology has been used to create models of individual people, such as Caryn Marjorie. However, these chatbots are not based on people's actual emotional and physiological responses to their environment, so they are, at best, surface-level approximations to the characters they are imitating. This paper describes how a new type of foundation model - a first-person foundation model - could be created from recordings of what a person sees and hears as well as their emotional and physiological reactions to these stimuli. A first-person foundation model would map environmental stimuli to a person's emotional and physiological states, and map a person's emotional and physiological states to their behavior. First-person foundation models have many exciting applications, including a new type of recommendation engine, personal assistants, generative adversarial networks, dating and recruitment. To obtain training data for a first-person foundation model, we have developed a recording rig that captures what the wearer is seeing and hearing as well as their emotional and physiological states. This novel source of data could help to address the shortage of new data for building the next generation of foundation models.
Data is crucial for evidence-based policymaking and enhancing public services, including those at the Ministry of Finance of the Republic of Indonesia. However, the complexity and dynamic nature of governmental financial data and regulations can hinder decision-making. This study investigates the potential of Large Language Models (LLMs) to address these challenges, focusing on Indonesia's financial data and regulations. While LLMs are effective in the financial sector, their use in the public sector in Indonesia is unexplored. This study undertakes an iterative process to develop KemenkeuGPT using the LangChain with Retrieval-Augmented Generation (RAG), prompt engineering and fine-tuning. The dataset from 2003 to 2023 was collected from the Ministry of Finance, Statistics Indonesia and the International Monetary Fund (IMF). Surveys and interviews with Ministry officials informed, enhanced and fine-tuned the model. We evaluated the model using human feedback, LLM-based evaluation and benchmarking. The model's accuracy improved from 35% to 61%, with correctness increasing from 48% to 64%. The Retrieval-Augmented Generation Assessment (RAGAS) framework showed that KemenkeuGPT achieved 44% correctness with 73% faithfulness, 40% precision and 60% recall, outperforming several other base models. An interview with an expert from the Ministry of Finance indicated that KemenkeuGPT has the potential to become an essential tool for decision-making. These results are expected to improve with continuous human feedback.
The integration of artificial intelligence into agricultural practices, specifically through Consultation on Intelligent Agricultural Machinery Management (CIAMM), has the potential to revolutionize efficiency and sustainability in farming. This paper introduces a novel approach that leverages large language models (LLMs), particularly GPT-4, combined with multi-round prompt engineering to enhance decision-making processes in agricultural machinery management. We systematically developed and refined prompts to guide the LLMs in generating precise and contextually relevant outputs. Our approach was evaluated using a manually curated dataset from various online sources, and performance was assessed with accuracy and GPT-4 Scores. Comparative experiments were conducted using LLama-2-70B, ChatGPT, and GPT-4 models, alongside baseline and state-of-the-art methods such as Chain of Thought (CoT) and Thought of Thought (ThoT). The results demonstrate that our method significantly outperforms these approaches, achieving higher accuracy and relevance in generated responses. This paper highlights the potential of advanced prompt engineering techniques in improving the robustness and applicability of AI in agricultural contexts.
Today's large language models (LLMs) can solve challenging question-answering tasks, and prompt engineering techniques, such as chain-of-thought (CoT), have gained attention for enhancing the explanation and correctness of outputs. However, many models and techniques tend to produce excessively verbose and lengthy answers, leading to issues with both conciseness and generation time. To address this, this paper analyzes the impact of output lengths on LLM inference pipelines by introducing and proposing novel metrics to evaluate the \textit{correct conciseness} of a model and related prompting techniques. Then, we examine the impact of controlling output length through a refined prompt engineering strategy, Constrained-CoT (CCoT), which encourages the model to produce more concise outputs. To better understand the effects of such a prompt, we also introduce two additional scores for analyzing the conciseness, measured in terms of redundancy and information flow in generated answers. Experiments on pretrained LLMs and multiple datasets demonstrate the benefits of the proposed metrics and the effectiveness of CCoT across different models.
Topic modeling is a technique for organizing and extracting themes from large collections of unstructured text. Non-negative matrix factorization (NMF) is a common unsupervised approach that decomposes a term frequency-inverse document frequency (TF-IDF) matrix to uncover latent topics and segment the dataset accordingly. While useful for highlighting patterns and clustering documents, NMF does not provide explicit topic labels, necessitating subject matter experts (SMEs) to assign labels manually. We present a methodology for automating topic labeling in documents clustered via NMF with automatic model determination (NMFk). By leveraging the output of NMFk and employing prompt engineering, we utilize large language models (LLMs) to generate accurate topic labels. Our case study on over 34,000 scientific abstracts on Knowledge Graphs demonstrates the effectiveness of our method in enhancing knowledge management and document organization.
With the rapid development of Text-to-Image (T2I) models, biases in human image generation against demographic social groups become a significant concern, impacting fairness and ethical standards in AI. Some researchers propose their methods to tackle with the issue. However, existing methods are designed for specific models with fixed prompts, limiting their adaptability to the fast-evolving models and diverse practical scenarios. Moreover, they neglect the impact of hallucinations, leading to discrepancies between expected and actual results. To address these issues, we introduce VersusDebias, a novel and universal debiasing framework for biases in arbitrary T2I models, consisting of an array generation (AG) module and an image generation (IG) module. The self-adaptive AG module generates specialized attribute arrays to post-process hallucinations and debias multiple attributes simultaneously. The IG module employs a small language model to modify prompts according to the arrays and drives the T2I model to generate debiased images, enabling zero-shot debiasing. Extensive experiments demonstrate VersusDebias's capability to debias any models across gender, race, and age simultaneously. In both zero-shot and few-shot scenarios, VersusDebias outperforms existing methods, showcasing its exceptional utility. Our work is accessible at https://github.com/VersusDebias/VersusDebias to ensure reproducibility and facilitate further research.
This study presents a comparative analysis of the a complex SQL benchmark, TPC-DS, with two existing text-to-SQL benchmarks, BIRD and Spider. Our findings reveal that TPC-DS queries exhibit a significantly higher level of structural complexity compared to the other two benchmarks. This underscores the need for more intricate benchmarks to simulate realistic scenarios effectively. To facilitate this comparison, we devised several measures of structural complexity and applied them across all three benchmarks. The results of this study can guide future research in the development of more sophisticated text-to-SQL benchmarks. We utilized 11 distinct Language Models (LLMs) to generate SQL queries based on the query descriptions provided by the TPC-DS benchmark. The prompt engineering process incorporated both the query description as outlined in the TPC-DS specification and the database schema of TPC-DS. Our findings indicate that the current state-of-the-art generative AI models fall short in generating accurate decision-making queries. We conducted a comparison of the generated queries with the TPC-DS gold standard queries using a series of fuzzy structure matching techniques based on query features. The results demonstrated that the accuracy of the generated queries is insufficient for practical real-world application.
The recently proposed Segment Anything Model (SAM) is a general tool for image segmentation, but it requires additional adaptation and careful fine-tuning for medical image segmentation, especially for small, irregularly-shaped, and boundary-ambiguous anatomical structures such as the knee cartilage that is of interest in this work. Repaired cartilage, after certain surgical procedures, exhibits imaging patterns unseen to pre-training, posing further challenges for using models like SAM with or without general-purpose fine-tuning. To address this, we propose a novel registration-based prompt engineering framework for medical image segmentation using SAM. This approach utilises established image registration algorithms to align the new image (to-be-segmented) and a small number of reference images, without requiring segmentation labels. The spatial transformations generated by registration align either the new image or pre-defined point-based prompts, before using them as input to SAM. This strategy, requiring as few as five reference images with defined point prompts, effectively prompts SAM for inference on new images, without needing any segmentation labels. Evaluation of MR images from patients who received cartilage stem cell therapy yielded Dice scores of 0.89, 0.87, 0.53, and 0.52 for segmenting femur, tibia, femoral- and tibial cartilages, respectively. This outperforms atlas-based label fusion and is comparable to supervised nnUNet, an upper-bound fair baseline in this application, both of which require full segmentation labels for reference samples. The codes are available at: https://github.com/chrissyinreallife/KneeSegmentWithSAM.git
The key to OOD detection has two aspects: generalized feature representation and precise category description. Recently, vision-language models such as CLIP provide significant advances in both two issues, but constructing precise category descriptions is still in its infancy due to the absence of unseen categories. This work introduces two hierarchical contexts, namely perceptual context and spurious context, to carefully describe the precise category boundary through automatic prompt tuning. Specifically, perceptual contexts perceive the inter-category difference (e.g., cats vs apples) for current classification tasks, while spurious contexts further identify spurious (similar but exactly not) OOD samples for every single category (e.g., cats vs panthers, apples vs peaches). The two contexts hierarchically construct the precise description for a certain category, which is, first roughly classifying a sample to the predicted category and then delicately identifying whether it is truly an ID sample or actually OOD. Moreover, the precise descriptions for those categories within the vision-language framework present a novel application: CATegory-EXtensible OOD detection (CATEX). One can efficiently extend the set of recognizable categories by simply merging the hierarchical contexts learned under different sub-task settings. And extensive experiments are conducted to demonstrate CATEX's effectiveness, robustness, and category-extensibility. For instance, CATEX consistently surpasses the rivals by a large margin with several protocols on the challenging ImageNet-1K dataset. In addition, we offer new insights on how to efficiently scale up the prompt engineering in vision-language models to recognize thousands of object categories, as well as how to incorporate large language models (like GPT-3) to boost zero-shot applications. Code is publicly available at https://github.com/alibaba/catex.
Different users find different images generated for the same prompt desirable. This gives rise to personalized image generation which involves creating images aligned with an individual's visual preference. Current generative models are, however, unpersonalized, as they are tuned to produce outputs that appeal to a broad audience. Using them to generate images aligned with individual users relies on iterative manual prompt engineering by the user which is inefficient and undesirable. We propose to personalize the image generation process by first capturing the generic preferences of the user in a one-time process by inviting them to comment on a small selection of images, explaining why they like or dislike each. Based on these comments, we infer a user's structured liked and disliked visual attributes, i.e., their visual preference, using a large language model. These attributes are used to guide a text-to-image model toward producing images that are tuned towards the individual user's visual preference. Through a series of user studies and large language model guided evaluations, we demonstrate that the proposed method results in generations that are well aligned with individual users' visual preferences.
We tackle the challenge of predicting models' Out-of-Distribution (OOD) performance using in-distribution (ID) measurements without requiring OOD data. Existing evaluations with "Effective Robustness", which use ID accuracy as an indicator of OOD accuracy, encounter limitations when models are trained with diverse supervision and distributions, such as class labels (Vision Models, VMs, on ImageNet) and textual descriptions (Visual-Language Models, VLMs, on LAION). VLMs often generalize better to OOD data than VMs despite having similar or lower ID performance. To improve the prediction of models' OOD performance from ID measurements, we introduce the Lowest Common Ancestor (LCA)-on-the-Line framework. This approach revisits the established concept of LCA distance, which measures the hierarchical distance between labels and predictions within a predefined class hierarchy, such as WordNet. We assess 75 models using ImageNet as the ID dataset and five significantly shifted OOD variants, uncovering a strong linear correlation between ID LCA distance and OOD top-1 accuracy. Our method provides a compelling alternative for understanding why VLMs tend to generalize better. Additionally, we propose a technique to construct a taxonomic hierarchy on any dataset using K-means clustering, demonstrating that LCA distance is robust to the constructed taxonomic hierarchy. Moreover, we demonstrate that aligning model predictions with class taxonomies, through soft labels or prompt engineering, can enhance model generalization. Open source code in our Project Page: https://elvishelvis.github.io/papers/lca/.
The rapid development of Large Language Models (LLMs) has brought impressive advancements across various tasks. However, despite these achievements, LLMs still pose inherent safety risks, especially in the context of jailbreak attacks. Most existing jailbreak methods follow an input-level manipulation paradigm to bypass safety mechanisms. Yet, as alignment techniques improve, such attacks are becoming increasingly detectable. In this work, we identify an underexplored threat vector: the model's internal reasoning process, which can be manipulated to elicit harmful outputs in a more stealthy way. To explore this overlooked attack surface, we propose a novel black-box jailbreak attack method, Analyzing-based Jailbreak (ABJ). ABJ comprises two independent attack paths: textual and visual reasoning attacks, which exploit the model's multimodal reasoning capabilities to bypass safety mechanisms, comprehensively exposing vulnerabilities in its reasoning chain. We conduct extensive experiments on ABJ across various open-source and closed-source LLMs, VLMs, and RLMs. In particular, ABJ achieves high attack success rate (ASR) (82.1% on GPT-4o-2024-11-20) with exceptional attack efficiency (AE) among all target models, showcasing its remarkable attack effectiveness, transferability, and efficiency. Our work reveals a new type of safety risk and highlights the urgent need to mitigate implicit vulnerabilities in the model's reasoning process.
The increasing volume of data in relational databases and the expertise needed for writing SQL queries pose challenges for users to access and analyze data. Text-to-SQL (Text2SQL) solves the issues by utilizing natural language processing (NLP) techniques to convert natural language into SQL queries. With the development of Large Language Models (LLMs), a range of LLM-based Text2SQL methods have emerged. This survey provides a comprehensive review of LLMs in Text2SQL tasks. We review benchmark datasets, prompt engineering methods, fine-tuning methods, and base models in LLM-based Text2SQL methods. We provide insights in each part and discuss future directions in this field.
We initiate a formal investigation into the design and analysis of LLM-based algorithms, i.e. algorithms that contain one or multiple calls of large language models (LLMs) as sub-routines and critically rely on the capabilities of LLMs. While LLM-based algorithms, ranging from basic LLM calls with prompt engineering to complicated LLM-powered agent systems and compound AI systems, have achieved remarkable empirical success, the design and optimization of them have mostly relied on heuristics and trial-and-errors, which is largely due to a lack of formal and analytical study for these algorithms. To fill this gap, we start by identifying the computational-graph representation of LLM-based algorithms, the design principle of task decomposition, and some key abstractions, which then facilitate our formal analysis for the accuracy and efficiency of LLM-based algorithms, despite the black-box nature of LLMs. Through extensive analytical and empirical investigation in a series of case studies, we demonstrate that the proposed framework is broadly applicable to a wide range of scenarios and diverse patterns of LLM-based algorithms, such as parallel, hierarchical and recursive task decomposition. Our proposed framework holds promise for advancing LLM-based algorithms, by revealing the reasons behind curious empirical phenomena, guiding the choices of hyperparameters, predicting the empirical performance of algorithms, and inspiring new algorithm design. To promote further study of LLM-based algorithms, we release our source code at https://github.com/modelscope/agentscope/tree/main/examples/paper_llm_based_algorithm.
Prompt engineering for large language models (LLMs) is often a manual time-intensive process that involves generating, evaluating, and refining prompts iteratively to ensure high-quality outputs. While there has been work on automating prompt engineering, the solutions generally are either tuned to specific tasks with given answers or are quite costly. We introduce GRAD-SUM, a scalable and flexible method for automatic prompt engineering that builds on gradient-based optimization techniques. Our approach incorporates user-defined task descriptions and evaluation criteria, and features a novel gradient summarization module to generalize feedback effectively. Our results demonstrate that GRAD-SUM consistently outperforms existing methods across various benchmarks, highlighting its versatility and effectiveness in automatic prompt optimization.
Prompt engineering is a crucial yet challenging task for optimizing the performance of large language models (LLMs) on customized tasks. This pioneering research introduces the Automatic Prompt Engineering Toolbox (APET), which enables GPT-4 to autonomously apply prompt engineering techniques. By leveraging sophisticated strategies such as Expert Prompting, Chain of Thought, and Tree of Thoughts, APET empowers GPT-4 to dynamically optimize prompts, resulting in substantial improvements in tasks like Word Sorting (4.4% increase) and Geometric Shapes (6.8% increase). Despite encountering challenges in complex tasks such as Checkmate in One (-14.8%), these findings demonstrate the transformative potential of APET in automating complex prompt optimization processes without the use of external data. Overall, this research represents a significant leap in AI development, presenting a robust framework for future innovations in autonomous AI systems and highlighting the ability of GPT-4 to bring prompt engineering theory to practice. It establishes a foundation for enhancing performance in complex task performance and broadening the practical applications of these techniques in real-world scenarios.
Large language models (LLMs) have shown remarkable performance on many different Natural Language Processing (NLP) tasks. Prompt engineering plays a key role in adding more to the already existing abilities of LLMs to achieve significant performance gains on various NLP tasks. Prompt engineering requires composing natural language instructions called prompts to elicit knowledge from LLMs in a structured way. Unlike previous state-of-the-art (SoTA) models, prompt engineering does not require extensive parameter re-training or fine-tuning based on the given NLP task and thus solely operates on the embedded knowledge of LLMs. Additionally, LLM enthusiasts can intelligently extract LLMs' knowledge through a basic natural language conversational exchange or prompt engineering, allowing more and more people even without deep mathematical machine learning background to experiment with LLMs. With prompt engineering gaining popularity in the last two years, researchers have come up with numerous engineering techniques around designing prompts to improve accuracy of information extraction from the LLMs. In this paper, we summarize different prompting techniques and club them together based on different NLP tasks that they have been used for. We further granularly highlight the performance of these prompting strategies on various datasets belonging to that NLP task, talk about the corresponding LLMs used, present a taxonomy diagram and discuss the possible SoTA for specific datasets. In total, we read and present a survey of 44 research papers which talk about 39 different prompting methods on 29 different NLP tasks of which most of them have been published in the last two years.
Large language models (LLMs) exhibit human-like intelligence, enabling them to simulate human behavior and support various applications that require both humanized communication and extensive knowledge reserves. Efforts are made to personify LLMs with special training data or hand-crafted prompts, while correspondingly faced with challenges such as insufficient data usage or rigid behavior patterns. Consequently, personified LLMs fail to capture personified knowledge or express persistent opinion. To fully unlock the potential of LLM personification, we propose PersLLM, a framework for better data construction and model tuning. For insufficient data usage, we incorporate strategies such as Chain-of-Thought prompting and anti-induction, improving the quality of data construction and capturing the personality experiences, knowledge, and thoughts more comprehensively. For rigid behavior patterns, we design the tuning process and introduce automated DPO to enhance the specificity and dynamism of the models' personalities, which leads to a more natural opinion communication. Both automated metrics and expert human evaluations demonstrate the effectiveness of our approach. Case studies in human-machine interactions and multi-agent systems further suggest potential application scenarios and future directions for LLM personification.