Research
Hypotheses, experiments and evaluation of the model’s results.
We are building our own language model
Our own intelligence.
Our own future.
At ORCNEIT Lab, we are developing our own language model, ORCNEITGPT. The work includes preparing data, training, evaluating results and building software tools.
The model is currently in training. The project page describes the tasks we want it to help with.
01
Company goal
We want to build and develop artificial intelligence technologies independently.
We are starting with a language model. Alongside it, we are developing the software and computing infrastructure needed for research and training.
This work should provide a foundation for future ORCNEIT projects: knowledge and tools that we can revisit, test and refine.
02
Company and projects
- 01
Company
What ORCNEIT does, and how research, software development and work with data fit together.
- 02
ORCNEITGPT
A detailed introduction to the model, the future assistant and the tools we intend to build.
- 03
Development
The current stage of work on ORCNEITGPT.
03 · Key project
ORCNEITGPT
We plan to build an assistant and tools for working with information around ORCNEITGPT.
Our plans include help with writing, programming, documents and research tasks. The project page explains how these capabilities could work together.
Building them requires both a language model and software for handling files, searching and checking results. The project description covers both areas.
04 · Now at ORCNEIT
ORCNEITGPT
- Status
- In progress
- Stage
- Model training
Work is currently focused on training the model. We are evaluating its language capabilities, the quality of its output and the stability of training.
This status refers to the model itself. Agentic functions, file handling and integrations are described on the project page as plans for future development.
05
ORCNEIT today
Engineering
Software for preparing data, training and testing the model.
Computing
Computing resources and data storage for research.
ORCNEITGPT
The project that brings this work together, from preparing data to evaluating the language model.
06
From experiment to result
A result only makes sense when you know how it was obtained.
- 01
Research
- 02
Experiment
- 03
Measurement
- 04
Implementation
A hypothesis needs to be tested, and the result evaluated. When working on the model, we need to understand what has changed and which questions remain open.
Not every intermediate experiment becomes a publication. A public account needs a clear description of the test and conclusions supported by its results.
We do not disclose information that could compromise infrastructure security or the company’s intellectual property.
07
Help us study language
We are developing ORCNEIT Contributors to help us study how people communicate today.
You can write an original text, invent a dialogue or explain what an expression means in a particular situation. We need both examples and the explanations behind them.
Materials are reviewed. Submitting a text does not mean it immediately becomes training data.
08
Contact ORCNEIT
Contact ORCNEIT about research, projects and collaboration. Our contact page lists the addresses for business enquiries, press and support.
contact@orcneitlab.comORCNEIT LLC
ООО «ОИЦ НИНТ»
Research · Engineering · Artificial Intelligence · Neural Engineering · Computing