qwen_training_data_miner_prototype

What problem does it solve?

This Skill automates the extraction of high-quality, domain-specific training examples from a large log file (012.txt) to create instruction-tuning datasets for training AI models like Gemma.

Core Features & Use Cases

• Data Mining: Scans a large text file (012.txt) for specific patterns related to defined knowledge domains (e.g., MPS scoring, WSP application).
• Example Extraction: Converts identified patterns into a structured instruction-tuning format, including input, output, and rationale.
• Quality Filtering: Applies strict criteria to ensure only high-quality, complete, and unambiguous examples are retained.
• Dataset Generation: Outputs a ready-to-use JSON dataset for model training, complete with metadata, recommended configurations, and autonomous execution instructions.
• Use Case: Automatically generate a dataset for training an AI to score tasks using the MPS methodology by mining historical decision data.

Quick Start

Use the qwen_training_data_miner_prototype skill to mine 012.txt for MPS scoring examples and generate a training dataset.