FolioStart free

Computer Science · Literature

Research papers on Reasoning in large language models

Recent and highly-cited academic work on reasoning in large language models, gathered from Semantic Scholar, CrossRef and OpenAlex.

Search all 200M+ papers on this topic, free →Or track new reasoning in large language models papers automatically as they publish

  1. LLaMA: Open and Efficient Foundation Language Models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, et al. · 2023 · arXiv (Cornell University) · 3,951 citations

    We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community.

  2. Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models

    Tiffany H. Kung, Morgan Cheatham, Arielle Medenilla, et al. · 2023 · PLOS Digital Health · 3,724 citations

    We evaluated the performance of a large language model called ChatGPT on the United States Medical Licensing Exam (USMLE), which consists of three exams: Step 1, Step 2CK, and Step 3. ChatGPT performed at or near the passing threshold for all three exams without any specialized training or reinforcement. Additionally, ChatGPT demonstrated a high level of concordance and insight in its explanations. These results suggest that large language models may have the potential to assist with medical education, and potentially, clinical decision-making.

  3. Large language models encode clinical knowledge

    Karan Singhal, Shekoofeh Azizi, Tao Tu, et al. · 2023 · Nature · 3,567 citations

    Abstract Large language models (LLMs) have demonstrated impressive capabilities, but the bar for clinical applications is high. Attempts to assess the clinical knowledge of models typically rely on automated evaluations based on limited benchmarks. Here, to address these limitations, we present MultiMedQA, a benchmark combining six existing medical question answering datasets spanning professional medicine, research and consumer queries and a new dataset of medical questions searched online, HealthSearchQA. We propose a human evaluation framework for model answers along multiple axes including factuality, comprehension, reasoning, possible harm and bias. In addition, we evaluate Pathways Lan

  4. PaLM: Scaling Language Modeling with Pathways

    Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, et al. · 2022 · arXiv (Cornell University) · 2,134 citations

    Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning resu

  6. A Survey of Large Language Models

    Wayne Xin Zhao, Kun Zhou, Junyi Li, et al. · 2026 · Frontiers of Computer Science · 1,475 citations

    Abstract The rapid evolution of large language models (LLMs) has driven a transformative shift in artificial intelligence (AI), reshaping both research paradigms and practical applications. Distinguished from their predecessors by unprecedented scale and advanced capabilities, LLMs necessitate new frameworks for understanding their development, behavior, and societal impact. This survey systematically reviews recent advancements in LLM techniques across four key dimensions: (1) pre-training methodologies, which establish core model capabilities through large-scale self-supervised training, architectural innovations, and data curation strategies; (2) post-training techniques, including superv

  7. Large Language Models are Zero-Shot Reasoners

    Takeshi Kojima, Shixiang Gu, Machel Reid, et al. · 2022 · arXiv (Cornell University) · 1,114 citations

    Pretrained large language models (LLMs) are widely used in many sub-fields of natural language processing (NLP) and generally known as excellent few-shot learners with task-specific exemplars. Notably, chain of thought (CoT) prompting, a recent technique for eliciting complex multi-step reasoning through step-by-step answer examples, achieved the state-of-the-art performances in arithmetics and symbolic reasoning, difficult system-2 tasks that do not follow the standard scaling laws for LLMs. While these successes are often attributed to LLMs' ability for few-shot learning, we show that LLMs are decent zero-shot reasoners by simply adding "Let's think step by step" before each answer. Experi

  8. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning

    Daya Guo, Dejian Yang, Haowei Zhang, et al. · 2025 · Nature · 853 citations

    Abstract General reasoning represents a long-standing and formidable challenge in artificial intelligence (AI). Recent breakthroughs, exemplified by large language models (LLMs) 1,2 and chain-of-thought (CoT) prompting 3 , have achieved considerable success on foundational reasoning tasks. However, this success is heavily contingent on extensive human-annotated demonstrations and the capabilities of models are still insufficient for more complex problems. Here we show that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning (RL), obviating the need for human-labelled reasoning trajectories. The proposed RL framework facilitates the emergent development of

  9. Self-Consistency Improves Chain of Thought Reasoning in Language Models

    Xuezhi Wang, Wei, Jason, Dale Schuurmans, et al. · 2022 · arXiv (Cornell University) · 711 citations

    Chain-of-thought prompting combined with pre-trained large language models has achieved encouraging results on complex reasoning tasks. In this paper, we propose a new decoding strategy, self-consistency, to replace the naive greedy decoding used in chain-of-thought prompting. It first samples a diverse set of reasoning paths instead of only taking the greedy one, and then selects the most consistent answer by marginalizing out the sampled reasoning paths. Self-consistency leverages the intuition that a complex reasoning problem typically admits multiple different ways of thinking leading to its unique correct answer. Our extensive empirical evaluation shows that self-consistency boosts the

  10. A survey on multimodal large language models

    Shukang Yin, Chaoyou Fu, Sirui Zhao, et al. · 2024 · National Science Review · 666 citations

    Recently, the multimodal large language model (MLLM) represented by GPT-4V has been a new rising research hotspot, which uses powerful large language models (LLMs) as a brain to perform multimodal tasks. The surprising emergent capabilities of the MLLM, such as writing stories based on images and optical character recognition-free math reasoning, are rare in traditional multimodal methods, suggesting a potential path to artificial general intelligence. To this end, both academia and industry have endeavored to develop MLLMs that can compete with or even outperform GPT-4V, pushing the limit of research at a surprising speed. In this paper, we aim to trace and summarize the recent progress of

  11. Graph of Thoughts: Solving Elaborate Problems with Large Language Models

    Maciej Besta, Nils Blach, Ales Kubicek, et al. · 2024 · Proceedings of the AAAI Conference on Artificial Intelligence · 454 citations

    We introduce Graph of Thoughts (GoT): a framework that advances prompting capabilities in large language models (LLMs) beyond those offered by paradigms such as Chain-of-Thought or Tree of Thoughts (ToT). The key idea and primary advantage of GoT is the ability to model the information generated by an LLM as an arbitrary graph, where units of information ("LLM thoughts") are vertices, and edges correspond to dependencies between these vertices. This approach enables combining arbitrary LLM thoughts into synergistic outcomes, distilling the essence of whole networks of thoughts, or enhancing thoughts using feedback loops. We illustrate that GoT offers advantages over state of the art on diffe

  13. Enhancing Chain of Thought Prompting in Large Language Models via Reasoning Patterns

    Yufeng Zhang, Xuepeng Wang, Lingxiang Wu, et al. · 2025 · Proceedings of the AAAI Conference on Artificial Intelligence · 11 citations

    Chain of Thought (CoT) prompting can encourage language models to engage in multi-step logical reasoning. The quality of the provided demonstrations significantly influences the success of downstream inference tasks. Current unsupervised CoT methods primarily select examples based on the semantics of the questions, which can introduce noise and lack interpretability. In this paper, we propose leveraging reasoning patterns to enhance CoT prompting effectiveness. Reasoning patterns represent the process by which language models arrive at their final results. By utilizing prior knowledge and prompt-based methods from large models, we first construct task-specific pattern sets. We then select di

  14. Length Instruction Fine-Tuning with Chain-of-Thought (LIFT-COT): Enhancing Length Control and Reasoning in Edge-Deployed Large Language Models

    Pinzhe Chen, Zhen Li · 2025 · Electronics · 4 citations

    This paper investigates the effectiveness of Chain-of-Thought (COT) reasoning in addressing length bias when deploying Large Language Models (LLMs) on the device side within wireless network environments. By conducting systematic experiments on two representative datasets, we perform a comparative evaluation of LLM performance under standard prompting versus COT-augmented prompting, using two benchmark tests. The study emphasizes how the COT strategy regulates response length and improves compliance with predefined maximum length (max_len) constraints. Experimental results demonstrate that COT reasoning significantly mitigates length bias, enhances output length control, and improves overall

  16. BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models

    Shuaitong Liu, Renjue Li, Lijia Yu, et al. · 2026 · Proceedings of the AAAI Conference on Artificial Intelligence · 1 citations

    Recent advances in Chain-of-Thought (CoT) prompting have substantially improved the reasoning capabilities of large language models (LLMs), but have also introduced their computational efficiency as a new attack surface. In this paper, we propose BadThink, the first backdoor attack designed to deliberately induce "overthinking" behavior in CoT-enabled LLMs while ensuring stealth. When activated by carefully crafted trigger prompts, BadThink manipulates the model to generate inflated reasoning traces—producing unnecessarily redundant thought processes while preserving the consistency of final outputs. This subtle attack vector creates a covert form of performance degradation that significantl

  17. Discourse Structure as an Interpretable Signal for Detecting Hallucinated Chain-of-Thought Reasoning in Large Language Models

    Boris Galitsky · 2026 · AI

    Large language models can generate fluent chain-of-thought (CoT) reasoning that appears coherent while exhibiting systematic distortions in evidence weighting and hypothesis comparison. This paper studies hallucinated CoT as a discourse-structural phenomenon, not only a factual one. We introduce a diagnostic reasoning benchmark with paired grounded and hallucinated explanations, where traces differ in how they organize evidence, alternatives, and defeaters. We extract discourse tree features that summarize evidence allocation, contrast preservation, commitment timing, and evidence integration, and combine them with the Joint Knowledge–Reasoning Hallucination Measure (JKRHM). Experiments on t

  18. Optimizing Chain-of-Thought Reasoning in Large Language Models for Complex Reasoning Tasks

    Amelia Moreno,, Julian Harper, Samuel Bennett · 2026 · Global Media and Social Sciences Research Journal

    The advent of large language models has fundamentally transformed the landscape of natural language processing, bringing unprecedented capabilities in text generation, summarization, and translation. However, these models often struggle with complex multi-step reasoning tasks, leading to the development of chain-of-thought prompting strategies. While chain-of-thought approaches significantly improve performance by eliciting intermediate reasoning steps, they are frequently constrained by inefficiencies such as error accumulation, high computational latency, and the generation of hallucinated reasoning paths. This paper presents a comprehensive framework for optimizing chain-of-thought reason

  19. Methods for Enhancing the Stability and Robustness of Chain-of-Thought Reasoning in Large Language Models

    Zhen Xu · 2026 · Journal of Computer Technology and Electronic Research

    Chain-of-Thought reasoning in large language models enhances the ability to solve complex tasks by generating intermediate steps; however, its stability and robustness face significant challenges. The reasoning process is susceptible to internal randomness, knowledge ambiguity, and error propagation, while existing architectures lack effective mechanisms for process constraint and regulation. This paper systematically analyzes the inherent vulnerabilities of such reasoning and proposes stability optimization methods, including structured iterative refinement, multi-granularity consistency constraints, and dynamic uncertainty regulation. It constructs a robustness enhancement framework encomp

  22. An Empirical Investigation of Adaptive Cognitive Guidance Strategy Optimization Driven by Hierarchical Chain‐of‐Thought Reasoning in Large Language Models

    Jiaxin Sun · 2026 · Engineering Reports

    ABSTRACT As large‐scale language models (LLMs) gain traction in education, the core challenge for intelligent tutoring has shifted from static knowledge transfer to dynamic cognitive companionship. This study proposes a novel adaptive cognitive tutoring strategy driven by a hierarchical chain‐of‐reasoning. The strategy constructs a three‐layer reasoning framework—strategic, tactical, and detailed—allowing complex problems to be decomposed into manageable cognitive units so learners can progressively deepen their understanding and problem‐solving. This study also designs a cognitive‐state diagnosis module that extracts and quantifies learners' multi‐dimensional features in r

  23. Empirical Analysis of Chain-of-Thought and Solver-Augmented Large Language Models for Deductive Reasoning

    Ya Wang, Raja Havish Seggoju, Adrian Paschke · 2026 · Neurosymbolic Artificial Intelligence

    Large language models (LLMs) have demonstrated remarkable capabilities in multistep reasoning, yet fundamental questions remain about their ability to perform reliable deductive reasoning. Two contrasting paradigms have emerged: chain-of-thought (CoT) prompting, which treats LLMs as self-contained reasoners, and solver-augmented approaches, which use LLMs as semantic parsers to translate problems into formal logic for symbolic execution. We present a systematic comparison of these approaches under strict single-pass evaluation across three established benchmarks (ProofWriter, PrOntoQA, and LogicalDeduction) and controlled synthetic datasets. Our findings reveal three key

Write your paper with these sources

Folio is the integrity-first research workspace: search 200M+ papers, save sources, and write with citations that format themselves. Free for students and researchers.

Start writing free →