Unverified50% confidenceFactExact time
Raising Temperature (e.g., to 1.0) activates low-probability words, making the output more creative but potentially less logical; lowering the temperature (e.g., below 0.3) lets high-probability tokens perform steadily, making the results more controllable.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
Unverified温度是一个0到2之间的超参数，控制回答的随机性与创造性，低温(0.1-0.3)适合编程数学客服，高温(0.8-1.5)适合头脑风暴写作80% similarVerifiedtemperature为0时退化为贪婪解码，每步选取概率最高的Token，但仅能保证单次模型版本内的确定性79% similarVerifiedThe temperature parameter scales the Logits vector, and when temperature is 0, it degenerates into greedy decoding.79% similarUnverified较高的温度参数会增加输出的随机性，在测试场景中往往需要将其调低以保证输出稳定性79% similarVerifiedThe Temperature parameter of large models typically ranges from 0 to 2; the lower the temperature, the more deterministic and conservative the output, while the higher the temperature, the more diverse and creative the output.78% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1116561API
curl https://kongchang.com/api/v1/knowledge/claims/1116561MCP
get_claim(id=1116561)