Introuducing Colox! Colox is a reasoning model fine tuned from LLaMA 2 created by me trained to mimic thinking like humans do to solve hard problems. It is free open source, and smarter than GPT o1. And it was made with around $25 unlike most companys that spend billions. :)
150 Epochs
This model went through 150 epoches to reason and think like humans do by doing a ton of thinking to give the answer.
Trained on 4.5K Patterns
This has been trained on 4.5K of human speech and thoughts.
Notes
Colox does not know it's identity, it thinks it is LLaMA made by Meta AI due to past training data from Meta AI and does not have any data about it's new identity. Plus, it may not think sometimes, that is an issue I am currently fixing.
