This dataset is a subset of a copy of Jackrong/GLM-5.1-Reasoning-1M-Cleaned
Each text value is a complete chat conversation in google/gemma-4-26B-A4B-it chat template with thinking:
<|turn>system
<|think|><turn|>
<|turn>user
{user_prompt}<turn|>
<|turn>model
<|channel>thought
{GLM_5.1_thinking}<channel|>{GLM_5.1_final_answer}<turn|>
