You could always use a local model to generate some text and then see how often local model text passes this test compared to text from some known human source. Testing with a variety of local models would be even better.
You could always use a local model to generate some text and then see how often local model text passes this test compared to text from some known human source. Testing with a variety of local models would be even better.
pubmed.ncbi.nlm.nih.gov/9884175/
pubmed.ncbi.nlm.nih.gov/9884175/
I’m kidding of course.
Also, how many people are joining signal groups after vetting the group itself to establish a chain of trust to the group’s founders?
If Nick Shirley himself created the group, would you know? How?
I’m kidding of course.
Also, how many people are joining signal groups after vetting the group itself to establish a chain of trust to the group’s founders?
If Nick Shirley himself created the group, would you know? How?