We present the Mu-SHROOM shared task which is focused on detecting hallucinations and other overgeneration mistakes in the output of instruction-tuned large language models (LLMs). Mu-SHROOM addresses general-purpose LLMs in 14 languages, and frames the hallucination detection problem as a span-labeling task. We received 2,618 submissions from 43 participating teams employing diverse methodologies. The large number of submissions underscores the interest of the community in hallucination detection. We present the results of the participating systems and conduct an empirical analysis to identify key factors contributing to strong performance in this task. We also emphasize relevant current challenges, notably the varying degree of hallucinations across languages and the high annotator disagreement when labeling hallucination spans.

Large language models (LLMs) perform well in medical QA, but their effectiveness in Japanese contexts is limited due to privacy constraints that prevent the use of commercial models like GPT-4 in clinical settings. As a result, recent efforts focus on instruction-tuning open-source LLMs, though the potential of combining them with retrieval-augmented generation (RAG) remains underexplored. To bridge this gap, we are the first to explore a knowledge graph-based (KG) RAG framework for Japanese medical QA small-scale open-source LLMs. Experimental results show that KG-based RAG has only a limited impact on Japanese medical QA using small-scale open-source LLMs. Further case studies reveal that the effectiveness of the RAG is sensitive to the quality and relevance of the external retrieved content. These findings offer valuable insights into the challenges and potential of applying RAG in Japanese medical QA, while also serving as a reference for other low-resource languages. Keywords: Japanese Medical Question Answering RAG Small-Scale LLMs Knowledge Graph. 1 Introduction Large language models (LLMs) have achieved remarkable performance in medical question answering (QA), even demonstrating the ability to pass medical licensing exams (e.g., the United States Medical Licensing Examination, USMLE) [19], which highlights their potential to understand complex medical knowledge. In particular, recent research [23,25,28] has explored the use of retrieval-augmented generation (RAG) [6] to incorporate external medical knowledge into LLMs, effectively mitigating "hallucination" issues [17,29,30] and further enhancing their applicability in medical QA tasks.

The reliance on translated or adapted datasets from English or multilingual resources introduces challenges regarding linguistic and cultural suitability. This study addresses the need for robust and culturally appropriate benchmarks by evaluating the quality of 17 commonly used Turkish benchmark datasets. Using a comprehensive framework that assesses six criteria, both human and LLM-judge annotators provide detailed evaluations to identify dataset strengths and shortcomings. Our results reveal that 70% of the benchmark datasets fail to meet our heuristic quality standards. The correctness of the usage of technical terms is the strongest criterion, but 85% of the criteria are not satisfied in the examined datasets. Although LLM judges demonstrate potential, they are less effective than human annotators, particularly in understanding cultural common sense knowledge and interpreting fluent, unambiguous text. GPT-4o has stronger labeling capabilities for grammatical and technical tasks, while Llama3.3-70B excels at correctness and cultural knowledge evaluation. Our findings emphasize the urgent need for more rigorous quality control in creating and adapting datasets for low-resource languages.

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.

Reinforcement learning has been shown to improve the performance of large language models. However, traditional approaches like RLHF or RLAIF treat the problem as single-step. As focus shifts toward more complex reasoning and agentic tasks, language models must take multiple steps of text generation, reasoning and environment interaction before generating a solution. We propose a synthetic data generation and RL methodology targeting multi-step optimization scenarios. This approach, called Step-Wise Reinforcement Learning (SWiRL), iteratively generates multi-step reasoning and tool use data, and then learns from that data. It employs a simple step-wise decomposition that breaks each multi-step trajectory into multiple sub-trajectories corresponding to each action by the original model. It then applies synthetic data filtering and RL optimization on these sub-trajectories. We evaluated SWiRL on a number of multi-step tool use, question answering, and mathematical reasoning tasks. Our experiments show that SWiRL outperforms baseline approaches by 21.5%, 12.3%, 14.8%, 11.1%, and 15.3% in relative accuracy on GSM8K, HotPotQA, CofCA, MuSiQue, and BeerQA, respectively. Excitingly, the approach exhibits generalization across tasks: for example, training only on HotPotQA (text question-answering) improves zero-shot performance on GSM8K (a math dataset) by a relative 16.9%.

Feature attribution (FA) methods are common post-hoc approaches that explain how Large Language Models (LLMs) make predictions. Accordingly, generating faithful attributions that reflect the actual inner behavior of the model is crucial. In this paper, we introduce Noiser, a perturbation-based FA method that imposes bounded noise on each input embedding and measures the robustness of the model against partially noised input to obtain the input attributions. Additionally, we propose an answerability metric that employs an instructed judge model to assess the extent to which highly scored tokens suffice to recover the predicted output. Through a comprehensive evaluation across six LLMs and three tasks, we demonstrate that Noiser consistently outperforms existing gradient-based, attention-based, and perturbation-based FA methods in terms of both faithfulness and answerability, making it a robust and effective approach for explaining language model predictions.

OpenAI announced today that users will soon be able to buy products through ChatGPT. The rollout of shopping buttons for AI-powered search queries will come to everyone, whether they are a signed-in user or not. Shoppers will not be able to check out inside of ChatGPT; instead they will be redirected to the merchant's website to finish the transaction. In a prelaunch demo for WIRED, Adam Fry, the ChatGPT search product lead at OpenAI, demonstrated how the updated user experience could be used to help people using the tool for product research decide which espresso machine or office chair to buy. The product recommendations shown to prospective shoppers are based on what ChatGPT remembers about a user's preferences as well as product reviews pulled from across the web.

In recent years, large language model (LLM)–based agents have garnered widespread attention across various fields. Their impressive capabilities, such as natural language communication,21,23 instruction following,26,28 and task execution,22,38 have the potential to expand both the format of information carriers and the way in which information is exchanged. LLM-based agents can now evolve into domain experts, becoming novel information carriers with domain-specific knowledge.1,28 For example, a Travel Agent can retain travel-related information within its parameters. LLM-based agents are also showcasing a new form of information exchange, facilitating more intuitive and natural interactions with users through dialogue and task execution.24,34 Figure 1 shows an example of these capabilities, in which users engage in dialogue with a Travel Agent to obtain information and complete their travel plans.

Copilot Arena is a VSCode extension that collects human preferences of code directly from developers. As model capabilities improve, large language models (LLMs) are increasingly integrated into user environments and workflows. In particular, software developers code with LLM-powered tools in integrated development environments such as VS Code, IntelliJ, or Eclipse. While these tools are increasingly used in practice, current LLM evaluations struggle to capture how users interact with these tools in real environments, as they are often limited to short user studies, only consider simple programming tasks as opposed to real-world systems, or rely on web-based platforms removed from development environments. To address these limitations, we introduce Copilot Arena, an app designed to evaluate LLMs in real-world settings by collecting preferences directly in a developer's actual workflow.

This dissertation establishes the contexture theory to mathematically characterize the mechanism of representation learning, or pretraining. Despite the remarkable empirical success of foundation models, it is not very clear what representations they learn, and why these representations are useful for various downstream tasks. A scientific understanding of representation learning is critical, especially at this point when scaling up the model size is producing diminishing returns, and designing new pretraining methods is imperative for further progress. Prior work treated different representation learning methods quite differently, whereas the contexture theory provides a unified framework for analyzing these methods. The central argument is that a representation is learned from the association between the input X and a context variable A. We prove that if an encoder captures the maximum information of this association, in which case we say that the encoder learns the contexture, then it will be optimal on the class of tasks that are compatible with the context. We also show that a context is the most useful when the association between X and A is neither too strong nor too weak. The important implication of the contexture theory is that increasing the model size alone will achieve diminishing returns, and further advancements require better contexts. We demonstrate that many pretraining objectives can learn the contexture, including supervised learning, self-supervised learning, generative models, etc. Then, we introduce two general objectives -- SVME and KISE, for learning the contexture. We also show how to mix multiple contexts together, an effortless way to create better contexts from existing ones. Then, we prove statistical learning bounds for representation learning. Finally, we discuss the effect of the data distribution shift from pretraining to the downstream task.