(无标题)
I ran a subset of simpleqa (150 questions, possibly fuzzy b/c i used rubric)
It would seem horizon alpha exhibits an interesting behavior of 'knowing when it doesn't know' more than other models https://t.co/YMFCxGaXFi
动态
I ran a subset of simpleqa (150 questions, possibly fuzzy b/c i used rubric)
It would seem horizon alpha exhibits an interesting behavior of 'knowing when it doesn't know' more than other models https://t.co/YMFCxGaXFi