LARGE LANGUAGE MODELS (LLMs)

Large Language Models (LLMs) are advanced artificial intelligence systems trained on massive volumes of textual data to understand, process and generate human-like language.They are based primarily on deep learning and transformer architecture, enabling them to perform tasks such as:

• Text generation
• Translation
• Summarisation
• Question answering
• Coding assistance
• Conversational interaction
• Reasoning and analysis

MAJOR LARGE LANGUAGE MODELS (LLMs) AND THEIR DEVELOPERS (2025–26)

NON-LLM AI SYSTEMS : UPSC SCIENCE & TECHNOLOGY NOTES

INTRODUCTION

Artificial Intelligence (AI) is much broader than Large Language Models (LLMs) such as GPT, Gemini or Claude. While LLMs focus mainly on understanding and generating human language, many AI systems operate without language modelling and are designed for specialised tasks such as image recognition, prediction, robotics, scientific discovery and decision-making.These systems are collectively referred to as Non-LLM AI Systems or Specialised AI Systems.

WHAT ARE NON-LLM AI SYSTEMS?

Non-LLM AI systems are artificial intelligence models that do not primarily rely on large-scale text generation.They generally focus on:

• Visual understanding
• Prediction and forecasting
• Robotics
• Scientific computation
• Decision optimisation
• Classification and anomaly detection

MAJOR TYPES OF NON-LLM AI SYSTEMS

1. COMPUTER VISION SYSTEMS

Meaning

AI systems designed to analyse and interpret images and videos.

Major Functions

• Object detection
• Image classification
• Facial recognition
• Video analytics
• Medical imaging

IMPORTANT EXAMPLES

2. NON-TEXT GENERATIVE AI MODELS

Meaning

AI systems capable of generating images, audio, proteins or structures instead of text.

IMPORTANT EXAMPLES

ALPHAFOLD : VERY IMPORTANT

AlphaFold revolutionised biology by accurately predicting 3D protein structures.

3. PREDICTIVE AND NUMERICAL AI

Meaning

AI systems used for forecasting future outcomes using numerical and statistical data.

IMPORTANT EXAMPLES

4. REINFORCEMENT LEARNING (RL) AGENTS

Meaning

AI systems that learn through trial-and-error interactions with environments to maximise rewards.

IMPORTANT EXAMPLES

5. ENCODER-ONLY LANGUAGE MODELS (NON-GENERATIVE)

Meaning

Models focused on language understanding rather than text generation.

IMPORTANT EXAMPLES

SMALL LANGUAGE MODELS (SLMs)

Meaning

Compact AI models designed to run efficiently on smaller hardware.

IMPORTANT EXAMPLES

IMPORTANT UPSC EXAMPLES