Practice
Learn
← Quit
0 / 0 correct
LLMs & AI
Hard
1 of 10
What does 'temperature 0' guarantee when calling an LLM API?
A
Greedy (most likely token) decoding, but not always bit-identical outputs, due to floating-point and batching nondeterminism
B
No hallucinations
C
Faster responses
D
Exactly identical outputs forever