reasoning-degeneration-dev/algorithmic-sft-countdown-distill-v2-partial
收藏Hugging Face2026-03-27 更新2026-03-29 收录
下载链接：
https://hf-mirror.com/datasets/reasoning-degeneration-dev/algorithmic-sft-countdown-distill-v2-partial
下载链接
链接失效反馈官方服务：
资源简介：
---
license: mit
tags:
- algorithmic-sft
- countdown
- distillation
- qwq-32b
- equation-answer
---
# algorithmic-sft-countdown-distill-v2-partial
Countdown QwQ-32B distillation with equation-answer format (partial: 5250 samples, 939 valid, 3940 invalid, 371 no_answer). Last batch: 5000.
## Dataset Info
- **Rows**: 5250
- **Columns**: 11
## Columns
| Column | Type | Description |
|--------|------|-------------|
| question | Value('string') | Full countdown prompt in equation-answer format |
| response | Value('string') | Raw QwQ-32B response (full, untruncated) |
| answer_block | Value('string') | Extracted content from <answer>...</answer> tags |
| status | Value('string') | Validation status: valid, invalid, or no_answer |
| error | Value('string') | Validation error message (empty if valid) |
| numbers | List(Value('int64')) | Available numbers for the countdown problem |
| target | Value('int64') | Target number to reach |
| difficulty | Value('int64') | *No description provided* |
| response_chars | Value('int64') | Character count of the full response |
| completion_tokens | Value('int64') | Number of tokens in the completion |
| finish_reason | Value('string') | vLLM finish reason (stop or length) |
## Generation Parameters
```json
{
"script_name": "scripts/collect_countdown_distill_v2.py",
"model": "Qwen/QwQ-32B",
"description": "Countdown QwQ-32B distillation with equation-answer format (partial: 5250 samples, 939 valid, 3940 invalid, 371 no_answer). Last batch: 5000.",
"hyperparameters": {
"temperature": 1.0,
"max_tokens": 32768,
"difficulty": 7
},
"input_datasets": []
}
```
## Usage
```python
from datasets import load_dataset
dataset = load_dataset("reasoning-degeneration-dev/algorithmic-sft-countdown-distill-v2-partial", split="train")
print(f"Loaded {len(dataset)} rows")
```
---
*This dataset is tracked in [reasoning-degeneration-dev/PROJECT-MANIFEST](https://huggingface.co/datasets/reasoning-degeneration-dev/PROJECT-MANIFEST)*
提供机构：
reasoning-degeneration-dev
