Having made many models that were only slightly better than a coin toss, that’s really not bad. Especially since that’s not even their primary design goal.
Comment on ChatGPT plays doctor with 72% success
Aceticon@lemmy.world 1 year ago
That’s a whole 22% better than a coin toss!!!
AlataOrange@lemmy.world 1 year ago
Rolder@reddthat.com 1 year ago
Not really comparable since a medical diagnosis has hundreds of possible results