Best of LLM — March 2024

1. 1
   Article

   Community Picks·2y

   OpenDevin/OpenDevin: 🐚 OpenDevin: Code Less, Make More

   OpenDevin is an open-source project aiming to replicate Devin, an autonomous AI software engineer. It allows users to run commands, generate scripts, and contribute to advancing software engineering with AI.

   262

   8
2. 2
   Article

   freeCodeCamp·2y

   How to Use Claude AI – Introduction to Claude AI + Code Example

   Claude AI is a leading large language model developed by Anthropic. It is safer and less likely to produce harmful and offensive output. Claude excels in conversation, content creation, language translation, visual processing, and code generation. Compared to ChatGPT, Claude has better factual accuracy, a larger context window, and a safer training approach.

   51

   4
3. 3
   Article

   GoPenAI·2y

   A Step-by-Step Guide to Creating a Large Language Model from scratch…

   This post provides a step-by-step guide to creating a Large Language Model (LLM) from scratch using the Transformer architecture and TensorFlow/Keras. It also explains how to implement transfer learning with Hugging Face.

   42
4. 4
   Article

   Python in Plain English·2y

   Build a Complete OpenSource LLM RAG QA Chatbot — An In-depth Journey (Introduction)

   Learn about the Retrieval-Augmented Generation (RAG) and its advantages and applications. Follow a comprehensive journey to build an OpenSource LLM RAG QA Chatbot.

   16
5. 5
   Article

   Hacker News·2y

   Skyvern-AI/skyvern: Automate browser-based workflows with LLMs and Computer Vision

   Automate browser-based workflows using LLMs and computer vision with Skyvern. It provides a simple API endpoint, is resistant to website layout changes, and leverages LLMs to cover complex situations. Skyvern Cloud offers managed automation at scale with anti-bot detection mechanisms, proxy network, and CAPTCHA solving.

   16
6. 6
   Article

   ITNEXT·2y

   Building My Own AI Workforce

   The author explores the possibility of two AI agents conversing and collaboratively creating the game 'tic tac toe'. The dialogue between the agents yielded unpredictable outcomes, with both successful and repetitive exchanges. Playing with an AI workforce demonstrates the potential for collaborative AI development.

   16
7. 7
   Article

   Medium·2y

   What are 1-bit LLMs?

   1-bit LLMs are a new addition to the Generative AI world that can help eliminate challenges associated with LLMs, especially their huge size. These LLMs use only 1 bit to store weight parameters, reducing the overall size and enabling use on smaller devices. BitNet b1.58, a 1-bit LLM variant, shows promising results in terms of performance and optimization.

   12

   1
8. 8
   Article

   GoDaddy Engineering·2y

   Everything I’ve Learned Building an AI App

   Learn the main concepts of AI and LLMs, understand how an LLM works, discover tips for making AI work better, and find out how to provide contextual information to the AI.

   11
9. 9
   Article

   KDnuggets·2y

   Getting Started With Claude 3 Opus That Just Destroyed GPT-4 and Gemini

   Anthropic has released a new series of large language models called Claude 3 that have outperformed GPT-4 and Gemini in benchmark tests. The models come in three variants: Haiku, Sonnet, and Opus. Claude 3 models have enhanced vision capabilities and can process various visual formats. To start using the latest Claude 3 model, you can create a new account on the Anthropic website.

   10
10. 10
    Article

    IEEE Spectrum·2y

    AI Prompt Engineering Is Dead

    Prompt engineering, the process of finding clever ways to phrase queries to large-language models (LLMs) or AI generators, may become obsolete as research suggests that models can optimize prompts themselves. Autotuned prompts have been successful in improving LLM performance on tasks like math questions and image generation. However, prompt engineering jobs will still be needed in the industry, albeit under different job titles. Adapting generative AI for industry requirements involves complex tasks that go beyond prompt engineering.

    10