Temperature
Randomness. Lower = more predictable.
↺
0
1.5
Top-P
Nucleus sampling threshold.
↺
0.1
1
Top-K
Limits candidates to top K tokens.
↺
1
100
Repetition penalty
Discourages repeated phrases.
↺
1
1.5
Max output tokens
Hard cap on response length.
↺
128
4096
Input — AI-generated text
Clear
Rewrite →
Output — rewritten text
Copy output