Gerard Sansai-cosmos.hashnode.dev·Nov 27, 2024Revealing the Hidden Threat: Algorithmic Bias in LLMsArtificial Intelligence (AI) is transforming industries and redefining how we interact with technology. From personalized recommendations to self-driving cars, AI promises efficiency and innovation. But beneath this progress lies a critical issue tha...35 readsAI
Gerard Sansai-cosmos.hashnode.dev·Nov 27, 2024How Fractals Illuminate the Nature of LLMsImagine a system that generates complexity from simple rules – like a snowflake forming its intricate pattern or a coastline revealing its mathematical precision at every scale. Large Language Models (LLMs) are revealing themselves as such a system: ...AI
Gerard Sansai-cosmos.hashnode.dev·Nov 21, 2024It's Time for AI Research to Embrace Evidence Over SpeculationThe current trajectory of AI research risks being undermined by a troubling pattern: the rush to adopt speculative methods and terminology without demanding the evidence to back them up. Techniques like in-context learning (ICL) and chain-of-thought ...83 readsAI
Gerard Sansai-cosmos.hashnode.dev·Nov 8, 2024Large Language Models Are Not Language Models but Stochastic Isomorphic EnginesIn recent years, the explosion of large language models (LLMs) like GPT has sparked a global conversation about machine "understanding" of language. Often described as "language models," these systems are credited with performing complex language tas...35 readsAI
Gareth Robertshyperpriors.io·Oct 31, 2024What AI Researchers Have Gotten So WrongIntroduction Explanation of the growing significance of AI in various fields The article explores the growing importance of AI across various domains, highlighting common misconceptions such as overestimating its capabilities and underestimating soc...AI Research
Gerard Sansai-cosmos.hashnode.dev·Oct 16, 2024The Impact of Knowledge Graphs on LLM Reasoning CapabilitiesAs AI models, particularly large language models (LLMs) like transformers, grow more sophisticated, the ability to support reasoning tasks has emerged as a critical frontier. While models like GPT have shown impressive abilities in pattern recognitio...27 readsAI
Gerard Sansai-cosmos.hashnode.dev·Oct 13, 2024Do Chatbots or LLMs Really Hallucinate? A Critical AnalysisIn recent years, the term "AI hallucinations" has become ubiquitous in discussions about the limitations and safety concerns of artificial intelligence. But what does this term really mean in the context of Large Language Models (LLMs)? This article ...31 readsAI
Gerard Sansai-cosmos.hashnode.dev·Oct 7, 2024Beyond Gradient Descent: Theoretical AnalysisAudience: ML Engineers, NLP Practitioners, AI Researchers Keywords: NLP, Transformers, Gradient Descent, Optimisation, Euclidean Geometry, Non-Euclidean Geometry, Hyperbolic Geometry, Elliptic Geometry, High-Dimensional Data, Discrete Data, Symbolic ...61 readsnlp transformers
Sai Chowdary Gullapallymlrad.io·Oct 6, 2024How OpenAI o1 models simulate human reasoningTLDR 🤏 The OpenAI o1 model uses a dedicated ‘reasoning stage’ to generate reasoning tokens, which help break down complex problems into manageable subtasks, simulating a structured thinking process similar to human reasoning. This dedicated reason...133 readsAI Reasoning
William Stetarcopin43.hashnode.dev·Sep 26, 2024Understanding Transformer Networks Through a Fluid Semiotic LensBy William Stetar Abstract: This article explores the concept of transformer-based language models operating under a "fluid semiotic regime." It delves into how these models process language dynamically, adjusting meanings based on context, and argu...Semiotics
