| dataset_info: | |
| features: | |
| - name: id | |
| dtype: string | |
| - name: conversations | |
| list: | |
| - name: content | |
| dtype: string | |
| - name: from | |
| dtype: string | |
| splits: | |
| - name: train | |
| num_bytes: 30583910 | |
| num_examples: 2959 | |
| download_size: 30538658 | |
| dataset_size: 30583910 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| Translating [Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled](https://huggingface.co/Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled)'s training dataset via TranslateGemma. | |
| The dataset consists of high-quality, filtered reasoning distillation data: | |
| | Dataset Name | Description / Purpose | | |
| |--------------|-----------------------| | |
| | [nohurry/Opus-4.6-Reasoning-3000x-filtered](https://huggingface.co/datasets/nohurry/Opus-4.6-Reasoning-3000x-filtered) | Provides comprehensive Claude 4.6 Opus reasoning trajectories. | | |
| | [Jackrong/Qwen3.5-reasoning-700x](https://huggingface.co/datasets/Jackrong/Qwen3.5-reasoning-700x) | Additional curated reasoning samples designed to strengthen structured step-by-step problem solving and improve reasoning diversity. | | |