Extracting Training Data from ChatGPT
Mozilla-Ocho/llamafile: Distribute and run LLMs with a single file.
Distribute and run LLMs with a single file. Contribute to Mozilla-Ocho/llamafile development by creating an account on GitHub.
Google Researchers’ Attack Prompts ChatGPT to Reveal Its Training Data
ChatGPT is full of sensitive private information and spits out verbatim text from CNN, Goodreads, WordPress blogs, fandom wikis, Terms of Service agreements, Stack Overflow source code, Wikipedia pages, news blogs, random internet comments, and much more.
God Help Us, Let's Try To Understand The Paper On AI Monosemanticity
Inside every AI is a bigger AI, trying to get out
Using llama.cpp with AWS instances · ggerganov/llama.cpp · Discussion #4225
Description The llama.cpp project offers unique ways of utilizing cloud computing resources. Here we will demonstrate how to deploy a llama.cpp server on a AWS instance for serving quantum and full...
Ekoparty 2023 - HACK THE PLANET - Prompt Injections in the Wild, real-world exploits and mitigations.
This talk sheds light on emerging attack techniques in AI applications such as Prompt Injections, a vulnerability at the very core of LLM Agents, Automated Tool Request Forgery, Data Exfiltration, and more.
Frost-group/The-Oracle-of-Zotero: LLM Chain querying a scientific Zotero library, with citations
LLM Chain querying a scientific Zotero library, with citations - Frost-group/The-Oracle-of-Zotero: LLM Chain querying a scientific Zotero library, with citations
Prompt Structure in Conversations with Generative AI
Most prompts contain a combination of the following components: request, framing context, format specification, and one or more references to previous answers or external sources. There are also 3 types of unconventional prompts that do not follow this structure: "Can you do X," "Give me more," and filler prompts.
Learnings from fine-tuning LLM on my Telegram messages
DALL·E image → GPT4 Vision → repeat | DALL·E Party
DALL·E image → GPT4 Vision → repeat
Prompt injection explained, November 2023 edition
A neat thing about podcast appearances is that, thanks to Whisper transcriptions, I can often repurpose parts of them as written content for my blog. One of the areas Nikita …
Generative AI for Beginners
Description
Fast Llama 2 on CPUs With Sparse Fine-Tuning and DeepSparse - Neural Magic
Key Takeaways We expanded our Sparse Fine-Tuning research results to include Llama 2. The results include 60% sparsity with INT8 quantization and no drop in accuracy. DeepSparse now supports accelerated inference of sparse-quantized Llama 2 models, with inference speeds 6-8x faster over the baseline at 60-80% sparsity. We used some interesting algorithmic techniques in order
Magnific AI
The most advanced AI upscaler & enhancer. Magnific can hallucinate and reimagine as many details as you wish guided by your own prompt and parameters!
visioncortex/vtracer: Raster to Vector Graphics Converter
Raster to Vector Graphics Converter. Contribute to visioncortex/vtracer development by creating an account on GitHub.
VectorArt.ai - Unlimited AI Generated Vector Images
The largest AI generated vector image library and AI vector art generator. Browse and download unlimited images for free or create your own with AI. SVGs and PNGs available for download. Prompt our specialized AI to generate a new unique image for your needs.
OpenAI researchers warned board of AI breakthrough ahead of CEO ouster, sources say
Ahead of OpenAI CEO Sam Altman’s four days in exile, several staff researchers wrote a letter to the board of directors warning of a powerful artificial intelligence discovery that they said could threaten humanity, two people familiar with the matter told Reuters.
Silicon Valley Is Turning Into Its Own Worst Fear
We asked a group of writers to consider the forces that have shaped our lives in 2017. Here, science fiction writer Ted Chiang looks at capitalism, Silicon Valley, and its fear of superintelligent AI.
simonw/ospeak: CLI tool for running text through OpenAI Text to speech
The AI industry turns against its favorite philosophy | Semafor
An ideology built on message boards is now struggling in the boardroom.
yl4579/StyleTTS2: StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models
StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models - yl4579/StyleTTS2: StyleTTS 2: Towards Human-Level Text-to-Speech ...
Is My Toddler a Stochastic Parrot?
The world is racing to develop ever more sophisticated large language models while a small language model unfurls itself in my home.
abi/screenshot-to-code: Drop in a screenshot and convert it to clean HTML/Tailwind/JS code
Drop in a screenshot and convert it to clean HTML/Tailwind/JS code - abi/screenshot-to-code: Drop in a screenshot and convert it to clean HTML/Tailwind/JS code
Death by AI - Survival Party Game
Step into a world where AI holds all the cards. Will it be your ally, or will it pull the plug? Choose wisely, because it's not programmed to care! Death by AI is a prompt based survival multiplayer game to play with your friends and family on mobile and web.
An early look at the possibilities as we experiment with AI and Music
Today we’re sharing a sneak peek at our first set of AI-related music experiments - Dream Track for Shorts and Music AI tools – built in collaboration with Google DeepMind.
Transforming the future of music creation
Announcing our most advanced music generation model and two new AI experiments, designed to open a new playground for creativity
Tiny LLMs
Powerful yet Compact: Browser-based AI models for a wide array of tasks, designed for optimal efficiency and user-friendliness, ensuring your privacy.
Oops! We Automated Bullshit. | Department of Computer Science and Technology
ChatGPT is a bullshit generator. To understand AI, we should think harder about bullshit
The 6 Types of Conversations with Generative AI
When interacting with generative-AI bots, users engage in six types of conversations, depending on their skill levels and their information needs. Interfaces for UI bots should support and accommodate this diversity of conversation styles.
The AI Pedagogy Project – metaLAB (at) Harvard