LLM large model training data (deep mathematical reasoning 860G) :

- The training data set for this LLM large model comes from a real-life mathematics case data set from teaching companies and universities . It is not AI-extracted data, and the levels range from elementary school to doctoral level. 90% of the mathematics problem materials contain full-process problem-solving methods, which is approximately 100-200 million questions.
- This LLM training data includes the original data set (such as real problem-solving pictures and materials), the post-processing data package (a format that can be directly used for large model training), and a user guide manual.
- The core of this LLM training data is the complete problem-solving process and explanation materials of mathematical data such as mathematics, physics, and statistics.
- All data in the documents are organized through classification, labeling, cleaning, format conversion, manual debugging, etc.
- The format contains codes, text, numbers, symbols, pictures, a small amount of audio/video, etc., 80% of which are post-processing code mode.

Purchase process:
- Choose to purchase the LLM large model training data package (Deep Mathematical Reasoning 860G) on the platform.
- After completing the purchase, the system will automatically send the data package download link to your email. You can directly download and extract it from the download link.
- If you purchase multiple data packages/amounts of large amounts (over HKD 100,000), you can place orders and pay in batches, or contact customer service to obtain large payment methods (such as corporate Alipay, corporate transfer, virtual currency, etc.).

Release date: March 7, 2025 (This data package will update and increase the data volume every 3 months. Users who have purchased it can get the latest data for free in the download link)
Upgrade: April 1, 2025, the second and third versions were launched. Each version has 0 repetition rate.

(Data Pack Usage Guide)

(Data Pack Directory - Topic Classification)

（Original data set - Teaching high-tech enterprise data）

(High-tech enterprise teaching materials - original data set example - complete problem-solving process, questions, answers)

(Large model training format - data example after post-processing)

(Final result - test result after training)