Model techniques map
Techniquesinference & servingreasoning control

implementation detail · filed under inference & serving

Think-tag response formatting

The response format uses <think> and </think> tokens to delimit reasoning content.

source
1
models
2
lab adopt it
1
strongest
used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

we utilize specialized response formats demarcated by the <think> and </think> tokens.

usedpost trainingin DeepSeek-V4DeepSeek

Filed alongside

Other methods under inference & serving :: reasoning control.