Integrating Retrieval Augmented Generation (RAG) with Large Language Models (LLMs) has shown the potential to provide precise, contextually relevant responses in knowledge intensive domains. This study investigates th e ap - plication of RAG for ABB circuit breakers, focusing on accuracy, reliability, and contextual relevance in high - stakes engineering environments. By leveraging tailored datasets, advanced embedding models, and optimized chunking strategies, the research addresses challenges in data retrieval and contextual alignment unique to engineering documentation. Key contributions include the development of a domain - specific dataset for ABB circuit breakers and the evaluation of three RAG pipelines: OpenAI GPT4o, C ohere, and Anthropic Claude. Advanced chunking methods, such as paragraph - based and title - aware segmentation, are assessed for their impact on retrieval accuracy and response generation. Results demonstrate that while certain configurations achieve high pr ecision and relevancy, limitations persist in ensuring factual faithfulness and completeness, critical in engineering contexts. This work underscores the need for iterative improvements in RAG systems to meet the stringent demands of electrical engineering tasks, including design, troubleshooting, and operational decision - making. The findings in this paper help advance research of AI in highly technical domains such as electrical engineering. Electrical engineering is a cornerstone of modern infrastructure, underpin n ing systems that power cities, enable communication, and drive technological innovation. From power generation and distribution to the design of advanced electronic systems, electrical engineering plays a vital role in ensuring the reliability, efficiency, and safety of critical infrastructure [1]. Mistakes or inaccuracies in the design, operation, or maintenance of e lectrical systems can have far - reaching consequences, including equipment failure, financial losses, and risks to public safety. In such high - stakes environments, precision and reliability in accessing accurate technical information are paramount [2]. Sim ilarly, in medicine, iterative retrieval methods have been proposed to enhance the accuracy of RAG systems. Xiong et al. [3] introduced the i - MedRAG system, which dynamically generates follow - up queries to refine responses. This approach improved retrieval accuracy and generalizability, although it incurred higher computational costs.

Concepts represent generalized abstractions that enable humans to categorize and reason efficiently, yet it is unclear to what extent Large Language Models (LLMs) comprehend these semantic relationships. Existing benchmarks typically focus on factual recall and isolated tasks, failing to evaluate the ability of LLMs to understand conceptual boundaries. To address this gap, we introduce CK-Arena, a multi-agent interaction game built upon the Undercover game, designed to evaluate the capacity of LLMs to reason with concepts in interactive settings. CK-Arena challenges models to describe, differentiate, and infer conceptual boundaries based on partial information, encouraging models to explore commonalities and distinctions between closely related concepts. By simulating real-world interaction, CK-Arena provides a scalable and realistic benchmark for assessing conceptual reasoning in dynamic environments. Experimental results show that LLMs' understanding of conceptual knowledge varies significantly across different categories and is not strictly aligned with parameter size or general model capabilities. The data and code are available at the project homepage: https://ck-arena.site.

Large Language Models (LLMs) have made substantial progress in recent years, yet evaluating their capabilities in practical Retrieval-Augmented Generation (RAG) scenarios remains challenging. In practical applications, LLMs must demonstrate complex reasoning, refuse to answer appropriately, provide precise citations, and effectively understand document layout. These capabilities are crucial for advanced task handling, uncertainty awareness, maintaining reliability, and structural understanding. While some of the prior works address these aspects individually, there is a need for a unified framework that evaluates them collectively in practical RAG scenarios. To address this, we present CReSt (A Comprehensive Benchmark for Retrieval-Augmented Generation with Complex Reasoning over Structured Documents), a benchmark designed to assess these key dimensions holistically. CReSt comprises 2,245 human-annotated examples in English and Korean, designed to capture practical RAG scenarios that require complex reasoning over structured documents. It also introduces a tailored evaluation methodology to comprehensively assess model performance in these critical areas. Our evaluation shows that even advanced LLMs struggle to perform consistently across these dimensions, underscoring key areas for improvement. We release CReSt to support further research and the development of more robust RAG systems. The dataset and code are available at: https://github.com/UpstageAI/CReSt.

End-to-end training of Spoken Language Models (SLMs) commonly involves adapting pre-trained text-based Large Language Models (LLMs) to the speech modality through multi-stage training on diverse tasks such as ASR, TTS and spoken question answering (SQA). Although this multi-stage continual learning equips LLMs with both speech understanding and generation capabilities, the substantial differences in task and data distributions across stages can lead to catastrophic forgetting, where previously acquired knowledge is lost. This paper investigates catastrophic forgetting and evaluates three mitigation strategies-model merging, discounting the LoRA scaling factor, and experience replay to balance knowledge retention with new learning. Results show that experience replay is the most effective, with further gains achieved by combining it with other methods. These findings provide insights for developing more robust and efficient SLM training pipelines.

Recent reasoning-oriented LLMs have demonstrated strong performance on challenging tasks such as mathematics and science examinations. However, core cognitive faculties of human intelligence, such as abstract reasoning and generalization, remain underexplored. To address this, we evaluate recent reasoning-oriented LLMs on the Abstraction and Reasoning Corpus (ARC) benchmark, which explicitly demands both faculties. We formulate ARC as a program synthesis task and propose nine candidate solvers. Experimental results show that repeated-sampling planning-aided code generation (RSPC) achieves the highest test accuracy and demonstrates consistent generalization across most LLMs. To further improve performance, we introduce an ARC solver, Knowledge Augmentation for Abstract Reasoning (KAAR), which encodes core knowledge priors within an ontology that classifies priors into three hierarchical levels based on their dependencies. KAAR progressively expands LLM reasoning capacity by gradually augmenting priors at each level, and invokes RSPC to generate candidate solutions after each augmentation stage. This stage-wise reasoning reduces interference from irrelevant priors and improves LLM performance. Empirical results show that KAAR maintains strong generalization and consistently outperforms non-augmented RSPC across all evaluated LLMs, achieving around 5% absolute gains and up to 64.52% relative improvement. Despite these achievements, ARC remains a challenging benchmark for reasoning-oriented LLMs, highlighting future avenues of progress in LLMs.

The ability to reason is one of the most fundamental capabilities of large language models (LLMs), enabling a wide range of downstream tasks through sophisticated problem-solving. A critical aspect of this is code reasoning, which involves logical reasoning with formal languages (i.e., programming code). In this paper, we enhance this capability of LLMs by exploring the following question: how can an LLM agent become progressively smarter in code reasoning with each solution it proposes, thereby achieving substantial cumulative improvement? Most existing research takes a static perspective, focusing on isolated problem-solving using frozen LLMs. In contrast, we adopt a cognitive-evolving perspective and propose a novel framework named Meta-Reflection with Cross-Referencing (MARCO) that enables the LLM to evolve dynamically during inference through self-improvement. From the perspective of human cognitive development, we leverage both knowledge accumulation and lesson sharing. In particular, to accumulate knowledge during problem-solving, we propose meta-reflection that reflects on the reasoning paths of the current problem to obtain knowledge and experience for future consideration. Moreover, to effectively utilize the lessons from other agents, we propose cross-referencing that incorporates the solution and feedback from other agents into the current problem-solving process. We conduct experiments across various datasets in code reasoning, and the results demonstrate the effectiveness of MARCO.

LLM-based digital twin simulation, where large language models are used to emulate individual human behavior, holds great promise for research in AI, social science, and digital experimentation. However, progress in this area has been hindered by the scarcity of real, individual-level datasets that are both large and publicly available. This lack of high-quality ground truth limits both the development and validation of digital twin methodologies. To address this gap, we introduce a large-scale, public dataset designed to capture a rich and holistic view of individual human behavior. We survey a representative sample of $N = 2,058$ participants (average 2.42 hours per person) in the US across four waves with 500 questions in total, covering a comprehensive battery of demographic, psychological, economic, personality, and cognitive measures, as well as replications of behavioral economics experiments and a pricing survey. The final wave repeats tasks from earlier waves to establish a test-retest accuracy baseline. Initial analyses suggest the data are of high quality and show promise for constructing digital twins that predict human behavior well at the individual and aggregate levels. By making the full dataset publicly available, we aim to establish a valuable testbed for the development and benchmarking of LLM-based persona simulations. Beyond LLM applications, due to its unique breadth and scale the dataset also enables broad social science research, including studies of cross-construct correlations and heterogeneous treatment effects.

When using supervised fine-tuning (SFT) to adapt large language models (LLMs) to specific domains, a significant challenge arises: should we use the entire SFT dataset for fine-tuning? Common practice often involves fine-tuning directly on the entire dataset due to limited information on the LLM's past training data. However, if the SFT dataset largely overlaps with the model's existing knowledge, the performance gains are minimal, leading to wasted computational resources. Identifying the unknown knowledge within the SFT dataset and using it to fine-tune the model could substantially improve the training efficiency. To address this challenge, we propose a self-learning framework for LLMs inspired by human learning pattern. This framework takes a fine-tuning (SFT) dataset in a specific domain as input. First, the LLMs answer the questions in the SFT dataset. The LLMs then objectively grade the responses and filter out the incorrectly answered QA pairs. Finally, we fine-tune the LLMs based on this filtered QA set. Experimental results in the fields of agriculture and medicine demonstrate that our method substantially reduces training time while achieving comparable improvements to those attained with full dataset fine-tuning. By concentrating on the unknown knowledge within the SFT dataset, our approach enhances the efficiency of fine-tuning LLMs.

Large language models (LLMs) have shown impressive performance by generating reasoning paths before final answers, but learning such a reasoning path requires costly human supervision. To address this issue, recent studies have explored self-training methods that improve reasoning capabilities using pseudo-labels generated by the LLMs themselves. Among these, confidence-based self-training fine-tunes LLMs to prefer reasoning paths with high-confidence answers, where confidence is estimated via majority voting. However, such methods exclusively focus on the quality of the final answer and may ignore the quality of the reasoning paths, as even an incorrect reasoning path leads to a correct answer by chance. Instead, we advocate the use of reasoning-level confidence to identify high-quality reasoning paths for self-training, supported by our empirical observations. We then propose a new self-training method, CORE-PO, that fine-tunes LLMs to prefer high-COnfidence REasoning paths through Policy Optimization. Our experiments show that CORE-PO improves the accuracy of outputs on four in-distribution and two out-of-distribution benchmarks, compared to existing self-training methods.

Large language models (LLMs) have demonstrated impressive capabilities in reasoning with the emergence of reasoning models like OpenAI-o1 and DeepSeek-R1. Recent research focuses on integrating reasoning capabilities into the realm of retrieval-augmented generation (RAG) via outcome-supervised reinforcement learning (RL) approaches, while the correctness of intermediate think-and-search steps is usually neglected. To address this issue, we design a process-level reward module to mitigate the unawareness of intermediate reasoning steps in outcome-level supervision without additional annotation. Grounded on this, we propose Learning to Think-and-Search (LeTS), a novel framework that hybridizes stepwise process reward and outcome-based reward to current RL methods for RAG. Extensive experiments demonstrate the generalization and inference efficiency of LeTS across various RAG benchmarks. In addition, these results reveal the potential of process- and outcome-level reward hybridization in boosting LLMs' reasoning ability via RL under other scenarios. The code will be released soon.