Artificial Intelligence & Machine Learning / Open-source AI models

Iterated Distillation and Amplification (IDA) is a training technique where the AI model internalizes the reasoning process through iterative policy improvement rather than relying on longer search times during inference. This method enables Deep Cogito v2 models to learn more efficient and accurate reasoning skills, improving performance on complex tasks such as math and language benchmarks while remaining cost-effective.

Deep Cogito v2 models are called hybrid reasoning models because they can toggle between two modes: a fast, direct-response mode for simple queries and a slower, step-by-step reasoning mode for complex problems. This hybrid approach allows the models to efficiently handle a wide range of tasks by balancing speed and depth of reasoning, outperforming other open-source models of similar size.

The Mixture of Experts (MoE) architecture means the AI model is designed like a board of 128 specialist advisors rather than a single monolithic system. When a query is received, the system selects the four most relevant experts to handle the task. This approach allows the model to have a very large number of parameters (120 billion) for vast knowledge while maintaining speed and efficiency by activating only a small part of the model at a time.

OpenAI's release of a powerful open-source AI model marks a return to its original open-source roots after years of guarding its top-tier models. This move is significant because it could reshape the competitive landscape by making advanced AI technology more accessible to developers and researchers, fostering innovation and collaboration in the AI community.