Large Language Models (LLMs) have emerged as powerful tools for generating coherent text, understanding context, and performing reasoning tasks. However, they struggle with temporal reasoning, which requires processing time-related information such as event sequencing, durations, and inter-temporal relationships. These capabilities are critical for applications including question answering, scheduling, and historical analysis. In this paper, we introduce TISER, a novel framework that enhances the temporal reasoning abilities of LLMs through a multi-stage process that combines timeline construction with iterative self-reflection. Our approach leverages test-time scaling to extend the length of reasoning traces, enabling models to capture complex temporal dependencies more effectively. This strategy not only boosts reasoning accuracy but also improves the traceability of the inference process. Experimental results demonstrate state-of-the-art performance across multiple benchmarks, including out-of-distribution test sets, and reveal that TISER enables smaller open-source models to surpass larger closed-weight models on challenging temporal reasoning tasks.

The growing need to integrate information from a large number of diverse sources poses significant scalability challenges for data integration systems. These systems often rely on manually written schema mappings, which are complex, source-specific, and costly to maintain as sources evolve. While recent advances suggest that large language models (LLMs) can assist in automating schema matching by leveraging both structural and natural language cues, key challenges remain. In this paper, we identify three core issues with using LLMs for schema mapping: (1) inconsistent outputs due to sensitivity to input phrasing and structure, which we propose methods to address through sampling and aggregation techniques; (2) the need for more expressive mappings (e.g., GLaV), which strain the limited context windows of LLMs; and (3) the computational cost of repeated LLM calls, which we propose to mitigate through strategies like data type prefiltering.

In this paper, we introduce CoRet, a dense retrieval model designed for code-editing tasks that integrates code semantics, repository structure, and call graph dependencies. The model focuses on retrieving relevant portions of a code repository based on natural language queries such as requests to implement new features or fix bugs. These retrieved code chunks can then be presented to a user or to a second code-editing model or agent. To train CoRet, we propose a loss function explicitly designed for repository-level retrieval. On SWE-bench and Long Code Arena's bug localisation datasets, we show that our model substantially improves retrieval recall by at least 15 percentage points over existing models, and ablate the design choices to show their importance in achieving these results.

Multimodal Large Language Models (MLLMs) have experienced rapid development in recent years. However, in the financial domain, there is a notable lack of effective and specialized multimodal evaluation datasets. To advance the development of MLLMs in the finance domain, we introduce FinMME, encompassing more than 11,000 high-quality financial research samples across 18 financial domains and 6 asset classes, featuring 10 major chart types and 21 subtypes. We ensure data quality through 20 annotators and carefully designed validation mechanisms. Additionally, we develop FinScore, an evaluation system incorporating hallucination penalties and multi-dimensional capability assessment to provide an unbiased evaluation. Extensive experimental results demonstrate that even state-of-the-art models like GPT-4o exhibit unsatisfactory performance on FinMME, highlighting its challenging nature. The benchmark exhibits high robustness with prediction variations under different prompts remaining below 1%, demonstrating superior reliability compared to existing datasets. Our dataset and evaluation protocol are available at https://huggingface.co/datasets/luojunyu/FinMME and https://github.com/luo-junyu/FinMME.

Most resources for evaluating social biases in Large Language Models are developed without co-design from the communities affected by these biases, and rarely involve participatory approaches. We introduce HESEIA, a dataset of 46,499 sentences created in a professional development course. The course involved 370 high-school teachers and 5,370 students from 189 Latin-American schools. Unlike existing benchmarks, HESEIA captures intersectional biases across multiple demographic axes and school subjects. It reflects local contexts through the lived experience and pedagogical expertise of educators. Teachers used minimal pairs to create sentences that express stereotypes relevant to their school subjects and communities. We show the dataset diversity in term of demographic axes represented and also in terms of the knowledge areas included. We demonstrate that the dataset contains more stereotypes unrecognized by current LLMs than previous datasets. HESEIA is available to support bias assessments grounded in educational communities.

Large language models (LLMs) have shown great potential in decision-making due to the vast amount of knowledge stored within the models. However, these pre-trained models are prone to lack reasoning abilities and are difficult to adapt to new environments, further hindering their application to complex real-world tasks. To address these challenges, inspired by the human cognitive process, we propose Causal-aware LLMs, which integrate the structural causal model (SCM) into the decision-making process to model, update, and utilize structured knowledge of the environment in a ``learning-adapting-acting" paradigm. Specifically, in the learning stage, we first utilize an LLM to extract the environment-specific causal entities and their causal relations to initialize a structured causal model of the environment. Subsequently,in the adapting stage, we update the structured causal model through external feedback about the environment, via an idea of causal intervention. Finally, in the acting stage, Causal-aware LLMs exploit structured causal knowledge for more efficient policy-making through the reinforcement learning agent. The above processes are performed iteratively to learn causal knowledge, ultimately enabling the causal-aware LLMs to achieve a more accurate understanding of the environment and make more efficient decisions. Experimental results across 22 diverse tasks within the open-world game ``Crafter" validate the effectiveness of our proposed method.

Aspect-Based Sentiment Analysis (ABSA) offers granular insights into opinions but often suffers from the scarcity of diverse, labeled datasets that reflect real-world conversational nuances. This paper presents an approach for generating synthetic ABSA data using Large Language Models (LLMs) to address this gap. We detail the generation process aimed at producing data with consistent topic and sentiment distributions across multiple domains using GPT-4o. The quality and utility of the generated data were evaluated by assessing the performance of three state-of-the-art LLMs (Gemini 1.5 Pro, Claude 3.5 Sonnet, and DeepSeek-R1) on topic and sentiment classification tasks. Our results demonstrate the effectiveness of the synthetic data, revealing distinct performance trade-offs among the models: DeepSeek-R1 showed higher precision, Gemini 1.5 Pro and Claude 3.5 Sonnet exhibited strong recall, and Gemini 1.5 Pro offered significantly faster inference. We conclude that LLM-based synthetic data generation is a viable and flexible method for creating valuable ABSA resources, facilitating research and model evaluation without reliance on limited or inaccessible real-world labeled data.

Byte Pair Encoding (BPE) tokenizers, widely used in Large Language Models, face challenges in multilingual settings, including penalization of non-Western scripts and the creation of tokens with partial UTF-8 sequences. Pretokenization, often reliant on complex regular expressions, can also introduce fragility and unexpected edge cases. We propose SCRIPT (Script Category Representation in PreTokenization), a novel encoding scheme that bypasses UTF-8 byte conversion by using initial tokens based on Unicode script and category properties. This approach enables a simple, rule-based pretokenization strategy that respects script boundaries, offering a robust alternative to pretokenization strategies based on regular expressions. We also introduce and validate a constrained BPE merging strategy that enforces character integrity, applicable to both SCRIPT-BPE and byte-based BPE. Our experiments demonstrate that SCRIPT-BPE achieves competitive compression while eliminating encoding-based penalties for non-Latin-script languages.

Large Language Models (LLMs) excel in various natural language processing tasks but remain vulnerable to generating harmful content or being exploited for malicious purposes. Although safety alignment datasets have been introduced to mitigate such risks through supervised fine-tuning (SFT), these datasets often lack comprehensive risk coverage. Most existing datasets focus primarily on lexical diversity while neglecting other critical dimensions. To address this limitation, we propose a novel analysis framework to systematically measure the risk coverage of alignment datasets across three essential dimensions: Lexical Diversity, Malicious Intent, and Jailbreak Tactics. We further introduce TRIDENT, an automated pipeline that leverages persona-based, zero-shot LLM generation to produce diverse and comprehensive instructions spanning these dimensions. Each harmful instruction is paired with an ethically aligned response, resulting in two datasets: TRIDENT-Core, comprising 26,311 examples, and TRIDENT-Edge, with 18,773 examples. Fine-tuning Llama 3.1-8B on TRIDENT-Edge demonstrates substantial improvements, achieving an average 14.29% reduction in Harm Score, and a 20% decrease in Attack Success Rate compared to the best-performing baseline model fine-tuned on the WildBreak dataset.

Cardiovascular disease (CVD) risk prediction models are essential for identifying high-risk individuals and guiding preventive actions. However, existing models struggle with the challenges of real-world clinical practice as they oversimplify patient profiles, rely on rigid input schemas, and are sensitive to distribution shifts. We developed AdaCVD, an adaptable CVD risk prediction framework built on large language models extensively fine-tuned on over half a million participants from the UK Biobank. In benchmark comparisons, AdaCVD surpasses established risk scores and standard machine learning approaches, achieving state-of-the-art performance. Crucially, for the first time, it addresses key clinical challenges across three dimensions: it flexibly incorporates comprehensive yet variable patient information; it seamlessly integrates both structured data and unstructured text; and it rapidly adapts to new patient populations using minimal additional data. In stratified analyses, it demonstrates robust performance across demographic, socioeconomic, and clinical subgroups, including underrepresented cohorts. AdaCVD offers a promising path toward more flexible, AI-driven clinical decision support tools suited to the realities of heterogeneous and dynamic healthcare environments.