haukurb

3 ms·

1. 1.▲
   by haukurb 2y ago

   The parent comment was making a philosophical argument, not a semantics one.
2. 2.▲
   by haukurb 3y ago

   It's the same definition from a talk by one of the PPO developers and also used elsewhere, i.e. it being first and foremost whether the output is inferred by applying proper epistemology (justified correct belief) to its training data.
3. 3.▲
   by haukurb 3y ago

   How does it compare to ranger, besides not being written in Python?
4. 4.▲
   by haukurb 3y ago

   Their example prompt is so bad that I'm split between them being wildly incompetent at understanding how transformers or autoregressiveness works (and general LLM dynamics) or they are deliberately obfuscating the task and prompt repre
5. 5.▲
   by haukurb 3y ago

   They might make it summarize intermittently behind the scenes when the context grows too large. Which would be how it has a gist of the stuff that was mentioned way earlier even thought it doesn't know the specifics.