Large language models (LLMs), adopted to understand human language, drive the development of artificial intelligence (AI) web search agents. Compared to traditional search engines, LLM-powered AI search agents are capable of understanding and responding to complex queries with greater depth, enabling more accurate operations and better context recognition. However, little attention and effort has been paid to the Chinese web search, which results in that the capabilities of open-source models have not been uniformly and fairly evaluated. The difficulty lies in lacking three aspects: an unified agent framework, an accurately labeled dataset, and a suitable evaluation metric. To address these issues, we propose a general-purpose and training-free web search agent by level-aware navigation, Level-Navi Agent, accompanied by a well-annotated dataset (Web24) and a suitable evaluation metric. Level-Navi Agent can think through complex user questions and conduct searches across various levels on the internet to gather information for questions. Meanwhile, we provide a comprehensive evaluation of state-of-the-art LLMs under fair settings. To further facilitate future research, source code is available at Github.

We review the recent literature (January 2022- October 2024) in South Asian languages on text-based language processing, multimodal models, and speech processing, and provide a spotlight analysis focused on 21 low-resource South Asian languages, namely Saraiki, Assamese, Balochi, Bhojpuri, Bodo, Burmese, Chhattisgarhi, Dhivehi, Gujarati, Kannada, Kashmiri, Konkani, Khasi, Malayalam, Meitei, Nepali, Odia, Pashto, Rajasthani, Sindhi, and Telugu. We identify trends, challenges, and future research directions, using a step-wise approach that incorporates relevance classification and clustering based on large language models (LLMs). Our goal is to provide a breadth-first overview of the recent developments in South Asian language technologies to NLP researchers interested in working with South Asian languages.

Human reliability analysis (HRA) is crucial for evaluating and improving the safety of complex systems. Recent efforts have focused on estimating human error probability (HEP), but existing methods often rely heavily on expert knowledge,which can be subjective and time-consuming. Inspired by the success of large language models (LLMs) in natural language processing, this paper introduces a novel two-stage framework for knowledge-driven reliability analysis, integrating IDHEAS and LLMs (KRAIL). This innovative framework enables the semi-automated computation of base HEP values. Additionally, knowledge graphs are utilized as a form of retrieval-augmented generation (RAG) for enhancing the framework' s capability to retrieve and process relevant data efficiently. Experiments are systematically conducted and evaluated on authoritative datasets of human reliability. The experimental results of the proposed methodology demonstrate its superior performance on base HEP estimation under partial information for reliability assessment.

The rapid identification of medical emergencies through digital communication channels remains a critical challenge in modern healthcare delivery, particularly with the increasing prevalence of telemedicine. This paper presents a novel approach leveraging large language models (LLMs) and prompt engineering techniques for automated emergency detection in medical communications. We developed and evaluated a comprehensive system using multiple LLaMA model variants (1B, 3B, and 7B parameters) to classify medical scenarios as emergency or non-emergency situations. Our methodology incorporated both system prompts and in-prompt training approaches, evaluated across different hardware configurations. The results demonstrate exceptional performance, with the LLaMA 2 (7B) model achieving 99.7% accuracy and the LLaMA 3.2 (3B) model reaching 99.6% accuracy with optimal prompt engineering. Through systematic testing of training examples within the prompts, we identified that including 10 example scenarios in the model prompts yielded optimal classification performance. Processing speeds varied significantly between platforms, ranging from 0.05 to 2.2 seconds per request. The system showed particular strength in minimizing high-risk false negatives in emergency scenarios, which is crucial for patient safety. The code implementation and evaluation framework are publicly available on GitHub, facilitating further research and development in this crucial area of healthcare technology.

AI is the workhorse of modern data analytics and omnipresent across many sectors. Large Language Models and multi-modal foundation models are today capable of generating code, charts, visualizations, etc. How will these massive developments of AI in data analytics shape future data visualizations and visual analytics workflows? What is the potential of AI to reshape methodology and design of future visual analytics applications? What will be our role as visualization researchers in the future? What are opportunities, open challenges and threats in the context of an increasingly powerful AI? This Visualization Viewpoint discusses these questions in the special context of biomedical data analytics as an example of a domain in which critical decisions are taken based on complex and sensitive data, with high requirements on transparency, efficiency, and reliability. We map recent trends and developments in AI on the elements of interactive visualization and visual analytics workflows and highlight the potential of AI to transform biomedical visualization as a research field. Given that agency and responsibility have to remain with human experts, we argue that it is helpful to keep the focus on human-centered workflows, and to use visual analytics as a tool for integrating ``AI-in-the-loop''. This is in contrast to the more traditional term ``human-in-the-loop'', which focuses on incorporating human expertise into AI-based systems.

Elucidating the rationale behind neural models' outputs has been challenging in the machine learning field, which is indeed applicable in this age of large language models (LLMs) and in-context learning (ICL). When it comes to estimating input attributions (IA), ICL poses a new issue of interpreting which example in the prompt, consisting of a set of examples, contributed to identifying the task/rule to be solved. To this end, in this paper, we introduce synthetic diagnostic tasks inspired by the poverty of the stimulus design in inductive reasoning; here, most in-context examples are ambiguous w.r.t. their underlying rule, and one critical example disambiguates the task demonstrated. The question is whether conventional IA methods can identify such an example in interpreting the inductive reasoning process in ICL. Our experiments provide several practical findings; for example, a certain simple IA method works the best, and the larger the model, the generally harder it is to interpret the ICL with gradient-based IA methods.

Recent generative large language models (LLMs) show remarkable performance in non-English languages, but when prompted in those languages they tend to express higher harmful social biases and toxicity levels. Prior work has shown that finetuning on specialized datasets can mitigate this behavior, and doing so in English can transfer to other languages. In this work, we investigate the impact of different finetuning methods on the model's bias and toxicity, but also on its ability to produce fluent and diverse text. Our results show that finetuning on curated non-harmful text is more effective for mitigating bias, and finetuning on direct preference optimization (DPO) datasets is more effective for mitigating toxicity. The mitigation caused by applying these methods in English also transfers to non-English languages. We find evidence that the extent to which transfer takes place can be predicted by the amount of data in a given language present in the model's pretraining data. However, this transfer of bias and toxicity mitigation often comes at the expense of decreased language generation ability in non-English languages, highlighting the importance of developing language-specific bias and toxicity mitigation methods.

Recent research in zero-shot Relation Extraction (RE) has focused on using Large Language Models (LLMs) due to their impressive zero-shot capabilities. However, current methods often perform suboptimally, mainly due to a lack of detailed, context-specific prompts needed for understanding various sentences and relations. To address this, we introduce the Self-Prompting framework, a novel method designed to fully harness the embedded RE knowledge within LLMs. Specifically, our framework employs a three-stage diversity approach to prompt LLMs, generating multiple synthetic samples that encapsulate specific relations from scratch. These generated samples act as in-context learning samples, offering explicit and context-specific guidance to efficiently prompt LLMs for RE. Experimental evaluations on benchmark datasets show our approach outperforms existing LLM-based zero-shot RE methods. Additionally, our experiments confirm the effectiveness of our generation pipeline in producing high-quality synthetic data that enhances performance.

Advancements in large language models (LLMs) have unlocked remarkable capabilities. While deploying these models typically requires server-grade GPUs and cloud-based inference, the recent emergence of smaller open-source models and increasingly powerful consumer devices have made on-device deployment practical. The web browser as a platform for on-device deployment is universally accessible, provides a natural agentic environment, and conveniently abstracts out the different backends from diverse device vendors. To address this opportunity, we introduce WebLLM, an open-source JavaScript framework that enables high-performance LLM inference entirely within web browsers. WebLLM provides an OpenAI-style API for seamless integration into web applications, and leverages WebGPU for efficient local GPU acceleration and WebAssembly for performant CPU computation. With machine learning compilers MLC-LLM and Apache TVM, WebLLM leverages optimized WebGPU kernels, overcoming the absence of performant WebGPU kernel libraries. Evaluations show that WebLLM can retain up to 80% native performance on the same device, with room to further close the gap. WebLLM paves the way for universally accessible, privacy-preserving, personalized, and locally powered LLM applications in web browsers. The code is available at: https://github.com/mlc-ai/web-llm.

Large Language Models (LLMs) aligned with human feedback have recently garnered significant attention. However, it remains vulnerable to jailbreak attacks, where adversaries manipulate prompts to induce harmful outputs. Exploring jailbreak attacks enables us to investigate the vulnerabilities of LLMs and further guides us in enhancing their security. Unfortunately, existing techniques mainly rely on handcrafted templates or generated-based optimization, posing challenges in scalability, efficiency and universality. To address these issues, we present JailPO, a novel black-box jailbreak framework to examine LLM alignment. For scalability and universality, JailPO meticulously trains attack models to automatically generate covert jailbreak prompts. Furthermore, we introduce a preference optimization-based attack method to enhance the jailbreak effectiveness, thereby improving efficiency. To analyze model vulnerabilities, we provide three flexible jailbreak patterns. Extensive experiments demonstrate that JailPO not only automates the attack process while maintaining effectiveness but also exhibits superior performance in efficiency, universality, and robustness against defenses compared to baselines. Additionally, our analysis of the three JailPO patterns reveals that attacks based on complex templates exhibit higher attack strength, whereas covert question transformations elicit riskier responses and are more likely to bypass defense mechanisms.