joelniklaus/dataforge-testing-20251215-111050

由[DataForge](https://github.com/huggingface/dataforge)生成的合成数据： * 模型：`LiquidAI/LFM2-350M` (`main`) * 源数据集：`simplescaling/s1K-1.1` (`train` 分割)。 * 生成配置：`temperature=1.0`, `top_p=1.0`, `top_k=50`, `max_tokens=8`, `model_max_context=2048` * 推测解码：`disabled` * 系统提示：`None` * 用户提示：列 `question` 该运行生成了1,000个样本和8,000个令牌。Synthetic data generated by [DataForge](https://github.com/huggingface/dataforge): * Model: `LiquidAI/LFM2-350M` (`main`) * Source dataset: `simplescaling/s1K-1.1` (`train` split). * Generation config: `temperature=1.0`, `top_p=1.0`, `top_k=50`, `max_tokens=8`, `model_max_context=2048` * Speculative decoding: `disabled` * System prompt: `None` * User prompt: Column `question` The run produced 1,000 samples and generated 8,000 tokens.