Task-oriented dialogue (TOD) systems facilitate goal-driven interactions between users and machines. While recent advances in deep learning have improved the performance, TOD systems often struggle in low-resource scenarios with limited labeled data. To address this challenge, we propose Spec-TOD, a novel framework designed to train an end-to-end TOD system with limited data. Spec-TOD introduces two main innovations: (i) a novel specialized end-to-end TOD framework that incorporates explicit task instructions for instruction-tuned large language models (LLMs), and (ii) an efficient training strategy that leverages lightweight, specialized LLMs to achieve strong performance with minimal supervision. Experiments on the MultiWOZ dataset, a widely used TOD benchmark, demonstrate that Spec-TOD achieves competitive results while significantly reducing the need for labeled data. These findings highlight the potential of the proposed framework in advancing efficient and effective TOD systems in low-resource settings.

The recent vaccine-related infodemic has amplified public concerns, highlighting the need for proactive misinformation management. We describe how we enhanced the reporting surveillance system of Victoria's vaccine safety service, SAEFVIC, through the incorporation of new information sources for public sentiment analysis, topics of discussion, and hesitancies about vaccinations online. Using VaxPulse, a multi-step framework, we integrate adverse events following immunisation (AEFI) with sentiment analysis, demonstrating the importance of contextualising public concerns. Additionally, we emphasise the need to address non-English languages to stratify concerns across ethno-lingual communities, providing valuable insights for vaccine uptake strategies and combating mis/disinformation. The framework is applied to real-world examples and a case study on women's vaccine hesitancy, showcasing its benefits and adaptability by identifying public opinion from online media.

Disseminators of disinformation often seek to attract attention or evoke emotions - typically to gain influence or generate revenue - resulting in distinctive rhetorical patterns that can be exploited by machine learning models. In this study, we explore linguistic and rhetorical features as proxies for distinguishing disinformative texts from other health and life-science text genres, applying both large language models and classical machine learning classifiers. Given the limitations of existing datasets, which mainly focus on fact checking misinformation, we introduce Four Shades of Life Sciences (FSoLS): a novel, labeled corpus of 2,603 texts on 14 life-science topics, retrieved from 17 diverse sources and classified into four categories of life science publications. The source code for replicating, and updating the dataset is available on GitHub: https://github.com/EvaSeidlmayer/FourShadesofLifeSciences

In this work, we introduce a novel approach for optimizing LLM training by adjusting learning rates across weights of different components in Transformer models. Traditional methods often apply a uniform learning rate across all network layers, potentially overlooking the unique dynamics of each part. Remarkably, our introduced relative learning rates, RLRS, method accelerates the training process by up to $23\%$, particularly in complex models such as Mixture of Experts (MoE). Hyperparameters of RLRS can be efficiently tuned on smaller models and then effectively reused on models up to $27\times$ larger. This simple and effective method results in a substantial reduction in training time and computational resources, offering a practical and scalable solution for optimizing large-scale neural networks.

Accident severity prediction plays a critical role in transportation safety systems but is a persistently difficult task due to incomplete data, strong feature dependencies, and severe class imbalance in which rare but high-severity cases are underrepresented and hard to detect. Existing methods often rely on monolithic models or black box prompting, which struggle to scale in noisy, real-world settings and offer limited interpretability. To address these challenges, we propose MARBLE a multiagent rule based LLM engine that decomposes the severity prediction task across a team of specialized reasoning agents, including an interchangeable ML-backed agent. Each agent focuses on a semantic subset of features (e.g., spatial, environmental, temporal), enabling scoped reasoning and modular prompting without the risk of prompt saturation. Predictions are coordinated through either rule-based or LLM-guided consensus mechanisms that account for class rarity and confidence dynamics. The system retains structured traces of agent-level reasoning and coordination outcomes, supporting in-depth interpretability and post-hoc performance diagnostics. Across both UK and US datasets, MARBLE consistently outperforms traditional machine learning classifiers and state-of-the-art (SOTA) prompt-based reasoning methods including Chain-of-Thought (CoT), Least-to-Most (L2M), and Tree-of-Thought (ToT) achieving nearly 90% accuracy where others plateau below 48%. This performance redefines the practical ceiling for accident severity classification under real world noise and extreme class imbalance. Our results position MARBLE as a generalizable and interpretable framework for reasoning under uncertainty in safety-critical applications.

While chain-of-thought (CoT) monitoring is an appealing AI safety defense, recent work on "unfaithfulness" has cast doubt on its reliability. These findings highlight an important failure mode, particularly when CoT acts as a post-hoc rationalization in applications like auditing for bias. However, for the distinct problem of runtime monitoring to prevent severe harm, we argue the key property is not faithfulness but monitorability. To this end, we introduce a conceptual framework distinguishing CoT-as-rationalization from CoT-as-computation. We expect that certain classes of severe harm will require complex, multi-step reasoning that necessitates CoT-as-computation. Replicating the experimental setups of prior work, we increase the difficulty of the bad behavior to enforce this necessity condition; this forces the model to expose its reasoning, making it monitorable. We then present methodology guidelines to stress-test CoT monitoring against deliberate evasion. Applying these guidelines, we find that models can learn to obscure their intentions, but only when given significant help, such as detailed human-written strategies or iterative optimization against the monitor. We conclude that, while not infallible, CoT monitoring offers a substantial layer of defense that requires active protection and continued stress-testing.

Large Language Models (LLMs) possess an extraordinary capability to produce text that is not only coherent and contextually relevant but also strikingly similar to human writing. They adapt to various styles and genres, producing content that is both grammatically correct and semantically meaningful. Recently, LLMs have been misused to create highly realistic phishing emails, spread fake news, generate code to automate cyber crime, and write fraudulent scientific articles. Additionally, in many real-world applications, the generated content including style and topic and the generator model are not known beforehand. The increasing prevalence and sophistication of artificial intelligence (AI)-generated texts have made their detection progressively more challenging. Various attempts have been made to distinguish machine-generated text from human-authored content using linguistic, statistical, machine learning, and ensemble-based approaches. This work focuses on two primary objectives Task-A, which involves distinguishing human-written text from machine-generated text, and Task-B, which attempts to identify the specific LLM model responsible for the generation. Both of these tasks are based on fine tuning of Generative Pre-trained Transformer (GPT_4o-mini), Large Language Model Meta AI (LLaMA) 3 8B, and Bidirectional Encoder Representations from Transformers (BERT). The fine-tuned version of GPT_4o-mini and the BERT model has achieved accuracies of 0.9547 for Task-A and 0.4698 for Task-B.

In-context learning enables transformer models to generalize to new tasks based solely on input prompts, without any need for weight updates. However, existing training paradigms typically rely on large, unstructured datasets that are costly to store, difficult to evaluate for quality and balance, and pose privacy and ethical concerns due to the inclusion of sensitive information. Motivated by these limitations and risks, we propose an alternative training strategy where we leverage a collection of multiple, small-scale, and domain-specific datasets. We empirically demonstrate that the increased quality and diversity of such data improve the generalization abilities of in-context learners beyond their training domain, while achieving comparable performance with models trained on a single large-scale dataset. We investigate this paradigm by leveraging meta-learning to train an in-context learner on the Meta-Album collection under several settings. Firstly, we show the performance in a controlled environment, where the test domain is completely excluded from the training knowledge. Secondly, we explore the robustness of these models to forgetting in a continual scenario where the information is accessible for a limited time. Finally, we explore the more challenging unsupervised scenario. Our findings demonstrate that transformers still generalize for in-context prediction when trained on a curated dataset collection while offering advantages in modularity and replaceability.

While Large Language Models (LLMs) demonstrate exceptional natural language capabilities, general-purpose models lack specialized domain knowledge for effective cybersecurity analysis. In this work, we investigate Domain-Adaptive Continuous Pretraining (DAP) as a methodology for enhancing cybersecurity understanding in pretrained LLMs while preserving general language capabilities. We systematically adapted three decoder-based architectures -- Llama-3.1-8B, DeepSeek-R1-Distill-Qwen-14B, and Llama-3.3-70B-Instruct -- using a curated 126-million-word cybersecurity corpus from standards, academic literature, and various other sources. Our approach employed constrained training parameters and distributed FSDP training to balance domain specialization with knowledge preservation. Evaluation across three cybersecurity benchmarks, namely, CTI-MCQ, CyberMetric, and SecEval, demonstrates consistent improvements post-adaptation. The Llama-3.3-70B-Ins-DAP model achieved state-of-the-art accuracies of 0.718, 0.933, and 0.864, respectively, outperforming specialized models, including Llama-Primus-Base. Notably, competitive performance was achieved using substantially smaller datasets (118.8 million versus 2.77 billion tokens), demonstrating efficient domain specialization viability. We establish that targeted continuous pretraining enables effective cybersecurity domain adaptation with computational feasibility, providing foundations for specialized AI assistants in threat analysis, vulnerability assessment, and security documentation while challenging prevailing assumptions about data requirements for LLM specialization.

This article presents and validates an ideal, four-stage workflow for the high-accuracy transcription and analysis of challenging medieval legal documents. The process begins with a specialized Handwritten Text Recognition (HTR) model, itself created using a novel "Clean Ground Truth" curation method where a Large Language Model (LLM) refines the training data. This HTR model provides a robust baseline transcription (Stage 1). In Stage 2, this baseline is fed, along with the original document image, to an LLM for multimodal post-correction, grounding the LLM's analysis and improving accuracy. The corrected, abbreviated text is then expanded into full, scholarly Latin using a prompt-guided LLM (Stage 3). A final LLM pass performs Named-Entity Correction (NEC), regularizing proper nouns and generating plausible alternatives for ambiguous readings (Stage 4). We validate this workflow through detailed case studies, achieving Word Error Rates (WER) in the range of 2-7% against scholarly ground truths. The results demonstrate that this hybrid, multi-stage approach effectively automates the most laborious aspects of transcription while producing a high-quality, analyzable output, representing a powerful and practical solution for the current technological landscape.