Synthetic QA & Instruction

The dataset consists of structured synthetic and human-verified question-answer pairs, initially sourced from Wikipedia-based QA tasks and later expanded to chatbot conversations and other AI tasks. Each entry includes the prompt, AI-generated response, and human evaluation scores. Aggregated over time, it supports training, benchmarking, and fine-tuning of large language models.