HN Reader

NewTopBestAskShowJob
LLMs show a "highly unreliable" capacity to describe own internal processes
score icon3
comment icon1
12 hours agoby pseudolus