Comment by curioussquirrel
11 hours ago
After Anthropic, Moonshot is another model provider who restricts tweaking of sampling parameters. I do like the idea of the vendor verifier, though.
11 hours ago
After Anthropic, Moonshot is another model provider who restricts tweaking of sampling parameters. I do like the idea of the vendor verifier, though.
What does "restricts tweaking of sampling parameters" mean?
"enforcing Temperature=1.0 and TopP=0.95 in Thinking mode, with mandatory validation that thinking content is correctly passed back."
If the post training is done with specific sampling parameters it would make sense to only use the parameters it was trained with.