Unverified50% confidenceFactExact time
合成数据通常由一个强大的教师模型生成问答对、推理链或代码样本，再用于训练学生模型
1
Sources
50%
Confidence
Long-term
Relevance
8/25/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe reasoning paradigm is essentially the application of reinforcement learning to the post-training stage of language models, allowing the model to autonomously explore reasoning paths and receive reward feedback based on answer correctness, first publicly demonstrated in the o1 series of models.74% similarUnverified大模型本质上是token生成器，训练目标是生成看起来自然的文本，而非精确的数据结构73% similarUnverified若训练集能充分覆盖推理时可能出现的工具类型和参数组合，小模型同样能学到可靠的格式生成能力，而不必依赖大模型的深层语义推理72% similarUnverifiedInstruction fine-tuning uses a dataset in question-answer pair format for training, enabling the model to better follow natural language instructions72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/797137API
curl https://kongchang.com/api/v1/knowledge/claims/797137MCP
get_claim(id=797137)