🚀 Medial Secures Investment on Shark Tank India - Fueling the Future of Professional Social Networking. 🔥

News

Messages

Try our Valuation Calculator →

Back

Rahul Agarwal

Founder | Agentic AI... • 2m

All LLMs are LMs, but not all LMs are LLMs. Most people still get confused. I've explained below. • 𝗟𝗠𝘀 (𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀): These are models that can process and generate human language. They can be small or medium-sized and may not require huge datasets. • 𝗟𝗟𝗠𝘀 (𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀): These are a 𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰 𝘁𝘆𝗽𝗲 of LM, but much 𝗹𝗮𝗿𝗴𝗲𝗿 in scale. LLMs like GPT-3 or GPT-4 are trained on massive datasets, have billions (or even trillions) of parameters. An 𝗟𝗟𝗠 is a 𝘀𝘂𝗯𝘀𝗲𝘁 𝗼𝗳 𝗟𝗠 that is: • Very large in size • Trained on massive datasets • Based on deep neural networks (Transformers) • Capable of reasoning, coding, summarizing, etc. Types of LM's: 𝗕𝘆 𝗦𝗶𝘇𝗲 / 𝗦𝗰𝗮𝗹𝗲 1. 𝗦𝗺𝗮𝗹𝗹 𝗟𝗠𝘀 • Lightweight, fast, low-cost models with limited intelligence. 2. 𝗠𝗲𝗱𝗶𝘂𝗺 𝗟𝗠𝘀 • Balanced speed and accuracy, suitable for most production systems. 3. 𝗟𝗮𝗿𝗴𝗲 𝗟𝗠𝘀 • High-capacity models with strong reasoning, powerful but expensive. ______________ 𝗕𝘆 𝗨𝘀𝗮𝗴𝗲 1. 𝗚𝗲𝗻𝗲𝗿𝗮𝗹-𝗽𝘂𝗿𝗽𝗼𝘀𝗲 𝗟𝗠𝘀 • Designed to handle many tasks • Chat, writing, coding, reasoning 2. 𝗗𝗼𝗺𝗮𝗶𝗻-𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰 𝗟𝗠𝘀 • Trained or tuned for one field • Legal, finance, medical, etc. • More accurate in narrow domains 3. 𝗘𝗱𝗴𝗲 𝗟𝗠𝘀 • Run locally on devices • Privacy-friendly • Limited power due to size ______________ 𝗕𝘆 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗦𝘁𝘆𝗹𝗲 1. 𝗣𝗿𝗲-𝘁𝗿𝗮𝗶𝗻𝗲𝗱 • Trained on general internet-scale data • Base intelligence layer 2. 𝗙𝗶𝗻𝗲-𝗧𝘂𝗻𝗲𝗱 • Adapted for specific tasks or domains • Improves accuracy and usefulness 3. 𝗜𝗻𝘀𝘁𝗿𝘂𝗰𝘁𝗶𝗼𝗻-𝘁𝘂𝗻𝗲𝗱 • Optimized to follow user instructions • This is what ChatGPT-style models are Most people just know about LLM's but it's important to know such fundamentals. ✅ Repost for others so they can also know this fundamental difference.

Recommendations from Medial

AI Engineer

AI Deep Explorer | f... • 10m

LLM Post-Training: A Deep Dive into Reasoning LLMs This survey paper provides an in-depth examination of post-training methodologies in Large Language Models (LLMs) focusing on improving reasoning capabilities. While LLMs achieve strong performance

Linkrcap Studio

A digital news platf... • 20h

India has no dearth of large language models (LLMs). Yet, most AI models struggle with the country itself 22 languages, multiple scripts, and patchy compute. Sarvam AI wants to fix this with its two indigenous models. But can it turn technical ambiti

Yash K

Avid Learner | In De... • 7d

India’s biggest AI drop so far 🇮🇳 Introducing Sarvam-30B and Sarvam-105B frontier LLMs built from India, for India. Sarvam-30B • 30B parameters (MoE) • 1B active params/token • 32K context window • Trained on 16T tokens • Competitive with Gemma-

AI Engineer

AI Deep Explorer | f... • 11m

"A Survey on Post-Training of Large Language Models" This paper systematically categorizes post-training into five major paradigms: 1. Fine-Tuning 2. Alignment 3. Reasoning Enhancement 4. Efficiency Optimization 5. Integration & Adaptation 1️⃣ Fin

Anonymous

Hey I am on Medial • 1y

Huge announcement from Meta. Welcome Llama 3.1🔥 This is all you need to know about it: The new models: - The Meta Llama 3.1 family of multilingual large language models (LLMs) is a collection of pre-trained and instruction-tuned generative models

1 Reply

Rahul Agarwal

Founder | Agentic AI... • 3m

SLM vs LLM — which AI model is best for you? I’ve explained both in simple steps below. 𝗦𝗟𝗠 (𝗦𝗺𝗮𝗹𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹) (𝘴𝘵𝘦𝘱-𝘣𝘺-𝘴𝘵𝘦𝘱) Lightweight AI models built for speed, focus, and on-device execution. 1. 𝗗𝗲𝗳𝗶𝗻𝗲

Shuvodip Ray
•

YouTube • 1y

Researchers at Google DeepMind introduced Semantica, an image-conditioned diffusion model capable of generating images based on the semantics of a conditioning image. The paper explores adapting image generative models to different datasets. Instea

2 Replies

Comet

#freelancer • 1y

Text Generation What It Is: Text generation involves using AI models to create humanlike text based on input prompts. How It Works: Models like GPT-3 use Transformer architectures. They’re pre-trained on vast text datasets to learn grammar, conte

1 Reply

Rahul Agarwal

Founder | Agentic AI... • 1m

Most people overlook these basics of AI Agents. I've explained it in a very simple way below. 1. 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 An AI system that observes its environment, information, makes decisions, and takes actions to achieve a goal. 2. 𝗟𝗟𝗠𝘀 (𝗟𝗮𝗿𝗴𝗲

Yogesh Dubey

Hey I am on Medial • 1y

Weekly AI Roundup : Cost Efficient Models, Advances in Robotics & Cutting edge AI tools OpenAI Unveils GPT-4o Mini: OpenAI's GPT-4o mini is a cost-efficient model aimed at expanding AI accessibility, offering a significant price reduction compared