Uncertainty quantification is a critical aspect of reinforcement learning and deep learning, with numerous applications ranging from efficient exploration and stable offline reinforcement learning to outlier detection in medical diagnostics. The scale of modern neural networks, however, complicates the use of many theoretically well-motivated approaches such as full Bayesian inference. Approximate methods like deep ensembles can provide reliable uncertainty estimates but still remain computationally expensive. In this work, we propose contextual similarity distillation, a novel approach that explicitly estimates the variance of an ensemble of deep neural networks with a single model, without ever learning or evaluating such an ensemble in the first place. Our method builds on the predictable learning dynamics of wide neural networks, governed by the neural tangent kernel, to derive an efficient approximation of the predictive variance of an infinite ensemble. Specifically, we reinterpret the computation of ensemble variance as a supervised regression problem with kernel similarities as regression targets. The resulting model can estimate predictive variance at inference time with a single forward pass, and can make use of unlabeled target-domain data or data augmentations to refine its uncertainty estimates. We empirically validate our method across a variety of out-of-distribution detection benchmarks and sparse-reward reinforcement learning environments. We find that our single-model method performs competitively and sometimes superior to ensemble-based baselines and serves as a reliable signal for efficient exploration. These results, we believe, position contextual similarity distillation as a principled and scalable alternative for uncertainty quantification in reinforcement learning and general deep learning.

Recent advancements in reasoning with large language models (RLLMs), such as OpenAI-O1 and DeepSeek-R1, have demonstrated their impressive capabilities in complex domains like mathematics and coding. A central factor in their success lies in the application of long chain-of-thought (Long CoT) characteristics, which enhance reasoning abilities and enable the solution of intricate problems. However, despite these developments, a comprehensive survey on Long CoT is still lacking, limiting our understanding of its distinctions from traditional short chain-of-thought (Short CoT) and complicating ongoing debates on issues like "overthinking" and "test-time scaling." This survey seeks to fill this gap by offering a unified perspective on Long CoT. (1) We first distinguish Long CoT from Short CoT and introduce a novel taxonomy to categorize current reasoning paradigms. (2) Next, we explore the key characteristics of Long CoT: deep reasoning, extensive exploration, and feasible reflection, which enable models to handle more complex tasks and produce more efficient, coherent outcomes compared to the shallower Short CoT. (3) We then investigate key phenomena such as the emergence of Long CoT with these characteristics, including overthinking, and test-time scaling, offering insights into how these processes manifest in practice. (4) Finally, we identify significant research gaps and highlight promising future directions, including the integration of multi-modal reasoning, efficiency improvements, and enhanced knowledge frameworks. By providing a structured overview, this survey aims to inspire future research and further the development of logical reasoning in artificial intelligence.

The rapid advancement of Large Multi-modal Models (LMMs) has enabled their application in scientific problem-solving, yet their fine-grained capabilities remain under-explored. In this paper, we introduce SciVerse, a multi-modal scientific evaluation benchmark to thoroughly assess LMMs across 5,735 test instances in five distinct versions. We aim to investigate three key dimensions of LMMs: scientific knowledge comprehension, multi-modal content interpretation, and Chain-of-Thought (CoT) reasoning. To unveil whether LMMs possess sufficient scientific expertise, we first transform each problem into three versions containing different levels of knowledge required for solving, i.e., Knowledge-free, -lite, and -rich. Then, to explore how LMMs interpret multi-modal scientific content, we annotate another two versions, i.e., Vision-rich and -only, marking more question information from texts to diagrams. Comparing the results of different versions, SciVerse systematically examines the professional knowledge stock and visual perception skills of LMMs in scientific domains. In addition, to rigorously assess CoT reasoning, we propose a new scientific CoT evaluation strategy, conducting a step-wise assessment on knowledge and logical errors in model outputs. Our extensive evaluation of different LMMs on SciVerse reveals critical limitations in their scientific proficiency and provides new insights into future developments. Project page: https://sciverse-cuhk.github.io

Time-series analysis is critical for a wide range of fields such as healthcare, finance, transportation, and energy, among many others. The practical applications often involve analyzing time-series data alongside contextual information in the form of natural language to support informed decisions. However, current time-series models are limited in their ability to perform reasoning that involves both time-series and their textual content. In this work, we address this gap by introducing \textit{Chat-TS}, a large language model (LLM) based framework, designed to support reasoning over time series and textual data. Unlike traditional models, Chat-TS integrates time-series tokens into LLMs' vocabulary, enhancing its reasoning ability over both modalities without compromising the core natural language capabilities, enabling practical analysis and reasoning across modalities. To support learning and evaluation in this setup, we contribute new datasets: the \textit{TS Instruct Training Dataset} which pairs diverse time-series data with relevant text instructions and responses for instruction tuning, the \textit{TS Instruct Question and Answer (QA) Gold Dataset} which provides multiple-choice questions designed to evaluate multimodal reasoning, and a \textit{TS Instruct Quantitative Probing Set} which contains a small subset of the TS Instruct QA tasks alongside math and decision-making questions for LLM evaluation. We designed a training strategy to preserve the inherent reasoning capabilities of LLMs while augmenting them for time-series reasoning. Experiments show that Chat-TS achieves state-of-the-art performance in multi-modal reasoning tasks by maintaining strong natural language proficiency while improving time-series reasoning. ~\footnote{To ensure replicability and facilitate future research, all models, datasets, and code will be available at [\texttt{Github-URL}].}

Evaluating mathematical capabilities is critical for assessing the overall performance of large language models (LLMs). However, existing evaluation methods often focus solely on final answers, resulting in highly inaccurate and uninterpretable evaluation outcomes, as well as their failure to assess proof or open-ended problems. To address these issues, we propose a novel mathematical process evaluation agent based on Tree-of-Error, called StepMathAgent. This agent incorporates four internal core operations: logical step segmentation, step scoring, score aggregation and error tree generation, along with four external extension modules: difficulty calibration, simplicity evaluation, completeness validation and format assessment. Furthermore, we introduce StepMathBench, a benchmark comprising 1,000 step-divided process evaluation instances, derived from 200 high-quality math problems grouped by problem type, subject category and difficulty level. Experiments on StepMathBench show that our proposed StepMathAgent outperforms all state-of-the-art methods, demonstrating human-aligned evaluation preferences and broad applicability to various scenarios. Our data and code are available at https://github.com/SHU-XUN/StepMathAgent.

Large Language Models (LLMs) are rapidly entering children's lives -- through parent-driven adoption, schools, and peer networks -- yet current AI ethics and safety research do not adequately address content-related risks specific to minors. In this paper, we highlight these gaps with a real-world case study of an LLMbased chatbot deployed in a middle school setting, revealing how students used and sometimes misused the system. We evaluate six prominent LLMs under different system prompts, demonstrating substantial variability in their childsafety compliance. Our results inform practical steps for more robust, childfocused safety mechanisms and underscore the urgency of tailoring AI systems to safeguard young users. Large Language Models (LLMs) have seen rapid adoption in educational settings, with both teachers and students recognizing their potential for personalized feedback and instant instructional support. Recent surveys indicate that over half of K-12 teachers in some regions now use LLMs for lesson planning, grading assistance, or creative class activities, while approximately onethird of students--some as young as 12--have experimented with such models for schoolwork (Common Sense Media, 2024). However, the emergence of LLMs in schools raises concerns about children's vulnerability. Children are still developing critical thinking skills, often place higher trust in authoritative-sounding answers, and may not fully understand an AI's limitations.

The performance of Large Language Models (LLMs) on multiple-choice question (MCQ) benchmarks is frequently cited as proof of their medical capabilities. We hypothesized that LLM performance on medical MCQs may in part be illusory and driven by factors beyond medical content knowledge and reasoning capabilities. To assess this, we created a novel benchmark of free-response questions with paired MCQs (FreeMedQA). Using this benchmark, we evaluated three state-of-the-art LLMs (GPT-4o, GPT-3.5, and LLama-3-70B-instruct) and found an average absolute deterioration of 39.43% in performance on free-response questions relative to multiple-choice (p = 1.3 * 10-5) which was greater than the human performance decline of 22.29%. To isolate the role of the MCQ format on performance, we performed a masking study, iteratively masking out parts of the question stem. At 100% masking, the average LLM multiple-choice performance was 6.70% greater than random chance (p = 0.002) with one LLM (GPT-4o) obtaining an accuracy of 37.34%. Notably, for all LLMs the free-response performance was near zero. Our results highlight the shortcomings in medical MCQ benchmarks for overestimating the capabilities of LLMs in medicine, and, broadly, the potential for improving both human and machine assessments using LLM-evaluated free-response questions.

Generative pretrained transformers (GPT) are the common large language models (LLMs) used for generating text from natural language inputs. However, the fixed properties of language parameters in individual LLMs can lead to inconsistencies in the generated outputs. This limitation also restricts the models' ability to represent diverse language patterns due to inherent biases. Moreover, many powerful LLMs are closed-source. This prevents organizations from integrating their data into these systems, raising concerns about data privacy and limiting industry applications. Inspired by the successful application of LLM ensemble models in text generation, recent literature has also investigated their potential in code generation. This article reviews these emerging LLM ensemble approaches. Our goal is to enhance readers' understanding of existing techniques and encourage further research and practical implementation, aiming to expand the real-world applications of LLM ensemble models in both text and code generation. We categorize these approaches into seven main methods: weight merging, knowledge fusion, mixture of experts, reward ensemble, output ensemble, routing, and cascading. From this list, we focus on four methods and models that show strong performance and potential for broader applications. We analyze their modeling steps, training methods, and output features to provide a clear understanding of their capabilities. Our findings highlight the benefits of LLM ensemble techniques. These include better representation of diversity, improved output quality, and greater flexibility in applications. This information offers valuable insights for selecting models for various real-world tasks involving text and code generation, and potentially applying methods to multimodal LLMs.

Causal graphs are commonly used to understand and model complex systems. Researchers often construct these graphs from different perspectives, leading to significant variations for the same problem. Comparing causal graphs is, therefore, essential for evaluating assumptions, integrating insights, and resolving disagreements. The rise of AI tools has further amplified this need, as they are increasingly used to generate hypothesized causal graphs by synthesizing information from various sources such as prior research and community inputs, providing the potential for automating and scaling causal modeling for complex systems. Similar to humans, these tools also produce inconsistent results across platforms, versions, and iterations. Despite its importance, research on causal graph comparison remains scarce. Existing methods often focus solely on structural similarities, assuming identical variable names, and fail to capture nuanced semantic relationships, which is essential for causal graph comparison. We address these gaps by investigating methods for comparing causal graphs from both semantic and structural perspectives. First, we reviewed over 40 existing metrics and, based on predefined criteria, selected nine for evaluation from two threads of machine learning: four semantic similarity metrics and five learning graph kernels. We discuss the usability of these metrics in simple examples to illustrate their strengths and limitations. We then generated a synthetic dataset of 2,000 causal graphs using generative AI based on a reference diagram. Our findings reveal that each metric captures a different aspect of similarity, highlighting the need to use multiple metrics.

The development of Large Language Models (LLMs) remains heavily skewed towards English and a few other high-resource languages. This linguistic disparity is particularly evident for Bangla - the 5th most spoken language. A few initiatives attempted to create open-source Bangla LLMs with performance still behind high-resource languages and limited reproducibility. To address this gap, we introduce TigerLLM - a family of Bangla LLMs. Our results demonstrate that these models surpass all open-source alternatives and also outperform larger proprietary models like GPT3.5 across standard benchmarks, establishing TigerLLM as the new baseline for future Bangla language modeling.