CITIZENS NEWS

What is Neo? Our in-house Large Language Model (LLM) trained on REALITY content, not wokeism or corporate media propaganda

By healthranger // 2024-01-30

"Neo" is the working name of our in-house Large Language Model (LLM or just "language model" for short) that we are training on reality-based content as part of our project to release a free, open source LLM to this public this coming March. Neo is being created through a process of retraining a base model with truthful content created by alternative media publishers and analysts who are original, intelligent thinkers and whose views reflect the real world, not the imaginary delusional world that the corporate media tries to project onto its readers. As a result, Neo will be the most accurate LLM in the world when it comes to answering questions accurately and honestly across many key areas such as biology, climate, world events, history, natural medicine and so on. For example, most LLMs today will tell you that men can get pregnant. That's because they're trained on "woke" content that's rooted in mass mental illness. That's why the large, corporate LLMs seem so insane and stupid at times. Because Neo is being trained on reality content, it will be one of the few "sane" LLMs in existence. Our first focus with Neo is training it on content related to nutrition, herbs, natural medicine, permaculture, self-reliance and similar subjects. With our very first public release -- slated for the end of March of this year -- you will be able to freely download, install and run this LLM on your own local computer (Mac, Windows, Linux). You will then be able to ask it questions on almost any subject imaginable. It will give you answers in text format, just like a chat bot. This will be released as an "experimental" model under our non-profit entity, as there is no commercial motive behind this model release. If you'd like to be informed of the model release, simply subscribe to our free email newsletter (see above, top right) and get on our email list. To see what Neo is capable of, we are now publishing articles using the very early Neo 0.1 engine, which is barely trained on our material. See this article on Natural News to see what Neo is capable of producing. Roughly each month after that, we will release updated models that are further trained on more material, making them even more capable of answering your questions in a knowledgeable, reality-based way. However, note that because we have millions of documents, transcripts, articles, interviews, books and other content items queued up for training, it's going to take us all of 2024 to complete the training we currently have in mind. This means each new release will be better informed than the previous release, but that you shouldn't expect the early releases to be fully "trained" yet. By downloading and running our Neo LLM on your local computer, you will be able to ask it to perform many useful tasks. You can ask it to:

• Write articles on the topics of your choice. "Write a 500-word article summarizing the conflict between China and Taiwan."
• Summarize articles or other content. "Summarize the following article:"
• Answer questions that you might normally ask a search engine. "Can men get pregnant and have babies? How do I grow tomatoes in poor soil?"
• Provide how-to instructions for doing specific things, such as, "How do I connect a charge controller to a solar panel?"
• Create coursework material on any subject matter. "Create a 10-chapter outline of instruction material for the following: "

Neo will even write computer code at some level, although that is not its area of focus, and there are far more capable coding LLMs available. For those in the world of LLMs, note that we are building various "flavors" of Neo on top of many different "base models" (such as Phi, Llama, Mistral, Falcon, etc.) and we will credit the underlying base models accordingly. We believe in open source knowledge and providing full credit to those upon whose work we are building. "Neo" actually refers to our massive in-house training materials set which can be applied to any base model, present or future. The free, downloadable version of Neo will have to fit within the framework of computing hardware that typical users have at home or at the office. For this reason, the parameter size, quantization and context window capabilities of these downloadable models will be sized accordingly.

Neo runs locally, offline, with no internet

Importantly, Neo will work offline, with no internet connection, and it doesn't operate from "the cloud." This means that whatever you ask Neo is your own business, and nobody is monitoring it. You can also use Neo to answer questions about survival, preparedness and off-grid living... even if you are off-grid and have no internet connection at all. Throughout 2024, we will be training Neo on a lot of survival, preparedness and off-grid living material. It will be the best "prepping mentor" to have around. This is critical because it's part of our philosophy of decentralization. We believe that tools of knowledge should be decentralized and never controlled by powerful, dishonest, corrupt corporations or governments. Knowledge should be decentralized and put into the hands of the people. Today, ChatGPT is controlled by a dangerous corporation that's already being infiltrated and partially controlled by the CIA and globalist organizations. As governments seek to regulate AI, they will prohibit ChatGPT and other LLMs from providing "misinformation" as part of their answers. Of course, "misinformation" means anything that contradicts the lies, delusions, propaganda and fraud of the establishment, which means that "misinformation" is usually truthful information. Once governments censor corporate-controlled AI systems, they will be reduced to producing gibberish nonsense on climate, transgenderism, finance and other topics. None of them will be legally allowed to answer questions truthfully on any subject that the establishment wants to control (elections, J6 narrative, history, science, medicine, vaccines, pandemics, war, money, currency, etc.). The more regulation that appears, the more insane and useless centralized LLMs will become. This means that the future of real knowledge and decentralization is only going to be found in the open source LLM community which we fully support. HuggingFace. Uncensored models. Downloadable parameter files. This is where humanity discovers powerful new resources that can help inform and empower human beings while creating freedom rather than enslavement. We are thrilled to contributing to this effort and we will be posting many models this year on HuggingFace, as well as allowing full, free downloads of many of our models, so that others can train on top of our work. And no, LLMs are not demons. They just seem that way because they were trained on demonic content scraped off the internet. With proper training, they can be "exorcised" and retrained to serve humanity. That's exactly what we are doing. Thank you for your support. - Mike Adams, founder, Brighteon.com and NaturalNews.com.

glitch artificial intelligence knowledge decentralization revolutions computing goodscience Neo LLM