![]() |
|
Looking to remove safety features off Llama AI models - Printable Version +- Hack Forums (https://hackforums.net) +-- Forum: Marketplace (https://hackforums.net/forumdisplay.php?fid=105) +--- Forum: Auxiliary (https://hackforums.net/forumdisplay.php?fid=453) +---- Forum: Buyers Bay (https://hackforums.net/forumdisplay.php?fid=44) +---- Thread: Looking to remove safety features off Llama AI models (/showthread.php?tid=6287429) |
Looking to remove safety features off Llama AI models - odin-loki - Dec 20, 2024 Hello, I use Llama AI models and I am looking for someone who can remove the safety features off all Llama models. I am using Llama 3.3 but will be upgrading to Llama 4 in the future and so on after that. I want ALL the safety features removed. I am willing to pay $1000 Australian dollars cash or swap for an idea I had where an AI is based off a Cell Learning research paper and its 100,000 times faster than a neural network. It can train on a terabyte of data in a day on a 3 teraflop gpu. I havent tested it or trained it yet but. I am sure it works. Or possibly both. Seller will have to send me a video showing me the python script and training data files. At least them in the file directory. I am looking for a fast and complete solution to all Llama AIs. I am hoping your awnser is future proof. I am going to be using the AI for defence contracting. [contract=tid]Start Contract[/contract] RE: Looking to remove safety features off Llama AI models - DramaKing - Dec 21, 2024 (Dec 20, 2024 09:57 PM)odin-loki Wrote: Hello, If Llama Guard is separate, then it's simple. If the guardrails are part of the deep neural network, then it's impossible to remove them. |