I did test the downloaded LLM and it worked perfectly on a computer that was not connected to the internet. Please be responsible and use the offline uncensored LLM for good rather than evil. Uncensored LLMs can be beneficial if used properly.
@GlobalScienceNetwork | gsnetwork.com
Summary of the key points from the video:
00:23 The video explains how to download a large language model (the Dolphin Llama 3 model) that was trained on the equivalent of 127 million novels or 2,500 times through Wikipedia. This model can be downloaded and run on an inexpensive external flash drive.
01:27 The video discusses how this uncensored, offline model provides access to information that may be censored or biased online. It allows for more privacy and the ability to interact with advanced AI without being monitored.
03:55 Offline language models can be used for proprietary, classified or personal information without the risk of that information being accessed by tech companies or governments.
06:43 The video provides step-by-step instructions for downloading and running the Dolphin Llama 3 model on an external drive, including using the Anything LLM interface for a more user-friendly experience.
13:30 GSN confirms the model is uncensored by testing it with a potentially sensitive question, and discusses other interfaces that can be used to run offline language models.
Uncensored LLM Info Thread
Related Videos
key points:
00:00 - The video discusses running a large 70 billion parameter language model (LLM) called Metaverse Llama 3 on an Intel Core Ultra processor with 96GB of RAM.
02:27 - The presenter walks through the steps to set up the Llama C++ library with the Intel IPEX LLM extension to run the model on the Intel GPU.
04:25 - The presenter tests running the 7 billion parameter Llama Q4 and Q5 models, seeing performance of around 10 tokens per second when offloading 33 out of 33 layers to the GPU.
08:10 - The presenter then downloads and runs the 70 billion parameter Metaverse Llama 3 model, which uses 94GB of RAM and 40GB of GPU memory, achieving around 1.4 tokens per second.
10:16 - The presenter notes that this setup allows running large LLMs on mobile Intel processors by leveraging the GPU, which was previously only possible on Apple Silicon.
key points:
00:00 - The video demonstrates how to download and install a local, private, and uncensored large language model (LLM) on a laptop, using models like DeepSEEK and Dolphin 3 LLaMA.
01:34 - There are major privacy and security concerns with using public LLMs like ChatGPT and DeepSEEK, as they may collect and store user data in China or other countries, exposing it to potential leaks and breaches.
03:27 - Running a local, open-source LLM avoids these privacy and security risks, and also allows for uncensored usage without the restrictions of paid or public LLMs.
04:16 - The video demonstrates the step-by-step process of downloading, installing, and using LLM Studio to run local instances of DeepSEEK and Dolphin 3 LLaMA models on the laptop.
08:02 - The uncensored Dolphin 3 LLaMA model allows for asking questions and receiving responses related to techniques for bypassing security software, which would not be possible with public, censored LLMs.