Category: Small Language Models

• DeepSeek’s AI Engine: A Look Under the Hood

  

  DeepSeek is making waves in the AI world by developing powerful and efficient AI models. The purpose of this post is to explore key concepts that explain how DeepSeek’s models learn, reason, and adapt.

  aitalks

  March 12, 2025

  AI, CoT, Distillation, MoE, Small Language Models

  CoT, deepseek, model distillation, MoE, Reinforcement Learning, supervised training, test-time scaling

• AI in 2024: A Year of Transformation and Technological Leaps

  

  2024 has been a landmark year for artificial intelligence. Unprecedented advancements have propelled AI into a driving force of societal impact and global transformation. This post explores the key trends defining this transformative AI landscape, highlighting its most significant advancements and challenges.

  aitalks

  December 23, 2024

  AI, AI Act, AI Agents, AI Challenges, AI Chip War, AI Competition, ChatGPT, Small Language Models, Training Data, year in review

  AI Agents, artificial intelligence, Ethical AI, gemini, Large Language Models (LLMs), OpenAI

• Unlocking the Potential of SLMs: A Look at Microsoft’s Orca 2

  

  Microsoft researchers are tackling the challenge of enhancing the reasoning abilities of small language models (SLMs). They demonstrate through Orca 2 that, despite operating with significantly fewer parameters than their large language model (LLM) counterparts, SLMs can offer a more accessible, efficient, and versatile AI solution when coupled with suitable training methods and optimized, synthetic…

  aitalks

  December 8, 2023

  AI, Small Language Models

  enhanced reasoning, Large Language Models (LLMs), Orca, Small Language Models (SLMs), student model, synthetic data