Temperature (Sampling)
A knob that controls AI randomness — low values produce deterministic output, high values produce creative/varied output.
Definition
Temperature scales the model's output probability distribution. T=0 = always pick the most likely token (deterministic). T=1.5 = wide variety. Most production uses T=0.7. Workers may adjust temperature when generating evaluation data.
Examples
- T=0 for math problems (need correct answer)
- T=0.9 for creative writing
Related terms
Looking for actual gigs that use Temperature (Sampling)? Browse all live listings or filter by $100+/hr.