Unverified50% confidenceFactExact time

温度是0到2之间的超参数，其底层机制是在Softmax函数前对logits进行缩放，温度趋近0时概率分布趋向one-hot，温度升高时分布趋于平坦

1

Sources

50%

Confidence

Long-term

Relevance

7/23/2026

First Seen

Sources

AI构建者手册：LLM原理、RAG架构与智能体设计全解

redditr/aiagents7/12/2026

Related Claims

UnverifiedTemperature scaling softens the prediction distribution by dividing by a learned temperature parameter T before the softmax, but its applicability to LLMs is limited by the unstructured nature of the output.74% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/599976

API

curl https://kongchang.com/api/v1/knowledge/claims/599976

MCP

get_claim(id=599976)