. . AI: Uncensored models & RAG (4.25.24)

"It's my computer, it should do what I want. My toaster toasts when I want. My car drives where I want. My lighter burns what I want. My knife cuts what I want. Why should the open-source AI running on my computer, get to decide for itself when it wants to answer my question?" So asked Eric Hartford in a blog post last year. This week, he and a team released an "uncensored" version of Llama-3-70b. Arguably also a matter of control, Cohere launched a toolkit for easy RAG (AI focused on your own documents).

-Marshall Kirkpatrick, Editor

First impacted: AI researchers, AI developers

The Cognitive Computations group, with members Eric Hartford, Lucas Atkins, and Fernando Fernandes, has launched Dolphin 2.9, an AI model based on Llama-3-70b. It's "uncensored" with allignment and bias/anti-bias instructions removed. As Hartford has written "Every demographic and interest group deserves their model. Open source is about letting people choose. The only way forward is composable alignment. To pretend otherwise is to prove yourself an idealogue and a dogmatist. There is no 'one true correct alignment' and even if there was, there's no reason why that should be OpenAI's brand of alignment." [cognitivecomputations/dolphin-2.9-llama3-70b · Hugging Face] Explore more of our coverage of: AI Model Development, Cognitive Computations, Meta's Llama-3 License. Share this story by email

First impacted: Web developers, Cloud engineers

Cohere has launched its Toolkit, a set of pre-built components that the company says enables users to swiftly construct and deploy RAG applications. The Toolkit, which includes a Next.js web app and pre-configured data sources, can be deployed on the Microsoft Azure Platform, run locally, or set up using AWS Sagemaker, Azure, or Cohere's platform, according to the company. [GitHub - cohere-ai/cohere-toolkit: Toolkit is a collection of prebuilt components enabling users to quickly build and deploy RAG applications.] Explore more of our coverage of: Cohere Toolkit, RAG Applications, Microsoft Azure Platform. Share this story by email