Tether Releases QVAC Genesis III Dataset to Enhance Small Model STEM Capabilities

Tether AI Research has released the synthetic dataset QVAC Genesis III, comprising 19.143 billion tokens across 159.6 million documents in 19 STEM fields, aimed at enhancing the capabilities of small AI models in science, technology, engineering, and mathematics. The dataset employs two methods: "failure analysis" and "option-level reasoning." The former involves a stronger teacher model analyzing the errors of the student model and providing correct answers, while the latter explains why the correct answer is correct and why other options are incorrect. Tests show that models trained on the complete corpus of Genesis III improved by 28.57, 21.35, and 15.03 percentage points on the ARC-Easy, ARC-Challenge, and MMLU STEM benchmark tests, respectively. The related paper has been accepted by the COLM 2026 conference.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.