4 ms·
There's nothing particular about this. This is just what you'd expect from a Language Model trained on large datasets. It reproduces a pattern commonly found in
by oulipo2 8mo ago
There's nothing particular about this. This is just what you'd expect from a Language Model trained on large datasets. It reproduces a pattern commonly found in documents
- alastairr 8mo agoNot sure I agree with you. For lower ability models, yes. Claude Opus 4.6 is incredibly capable, so it's odd to me it has this residual 'misspeak' behaviour.
- oulipo2 8mo agoThat's the issue, people are anthropomorphizing those models, but... they're all the same (conceptually). They just do random hallucinations, trying to make those hallucination match the "reality" (of their training data) as much as possible