Interesting, I had never heard about min-p until now. From what I understand, it's like a low-pass filter for the token sampling pool which boosts semantic coherence. Like removing static from the radio.
Do you have any benchmarks of min-p sampling with the new reasoning models, such as QwQ and R1?
Do you have any benchmarks of min-p sampling with the new reasoning models, such as QwQ and R1?