DataGenn AI Research:
Building LLMs and domain-specific language models for the AI era of innovation & discovery.
One-click deploy of pretrained foundation models or train DataGenn AI language models on your organization’s internal datasets.
Our base LLM is an ~80BP, Llama 2 model under the hood, trained further on critical datasets to solve the reversal curse, improve logic + reasoning abilities.
DataGenn.AI is currently developing our first domain specific LLM and plan to launch our invest model in the next 2 weeks (ships first week of November).
DataGenn Invest LLM is an investor and trader financial markets large language model that has been finely tuned to generate insightful investing / trading knowledge via conversational AI platforms.
Once developed upon by ML Engineers, the invest model can invest autonomously as an AI & Data powered trading bot that’s been trained and fine-tuned on critical financial markets data, logic and financial markets insights gained over 2 decades of investing, trading, and 6 years of active short selling Crypto, mostly shorting bitcoin (BTC).
DataGenn LLM is our large language model currently in development on the scope of ~80 billion parameters. DataGenn LLM is ideal for synthesizing complex information, content generation & engaging in human-like conversations through chat. It uses Meta’s Llama 2 LLM at its core and we’ve scaled it up from there. Sign up for early access beginning in November 2023.
Domain-Specific Language Models
With our investor trader pretrained language model, DataGenn Invest LLM is ~94 billion parameters. This is after further training and fine tuning Llama 2 70B.
At DataGenn.ai you can also train domain-specific language models on your own custom datasets for optimal output with low operating expenses structured.
Generative AI Development
DataGenn Development plans, develops & ships custom generative AI applications for startups, governments & enterprise organizations. You can also build your own tech on DataGenn LLMs to generate your imagination and innovate in the fast-paced generative AI marketplace.
Large Language Models uniquely trained with both RLHF & RLAIF training methods.
DataGenn.ai synthesizes data across multiple domains by using a combination of Reinforcement Learning from Human Feedback (RLHF) and Reinforcement Learning from AI Feedback (RLAIF).
This unique approach blending training methodologies is polymathic in nature, drawing conclusions & unique insights from separated domains of knowledge.
Our ultimate goal is to make DataGenn LLM a self-learning language model that can learn on its own, eventually recreating itself, iterating and self improving by design.