Yi 9B 200K

About

The Yi 9B 200K model by 01.AI is a bilingual AI model designed for English and Chinese language tasks, inspired by the Llama architecture but independently trained. It utilizes a vast 3 terabyte multilingual corpus and supports a generous context window of 200,000 tokens. This model excels in language understanding, commonsense reasoning, coding, and mathematical tasks, boasting robust performance in these areas. Though as a base model, it has limitations in instruction adherence and may encounter issues like hallucinations and inconsistent outputs, particularly in complex reasoning. Performance enhancements can be achieved by fine-tuning parameters such as temperature, top_p, and top_k 124.

Capabilities