News for hackers. Stuff that matters.
Re:Show HN: Llama 3.3 70B Sparse Autoencoders with API access by wg0 on Monday December 23, 2024 @03:27PM (#42497326)
Noob question - how do we know that these autoencoders aren't hallucinating and really are mapping/clustering what they should be?
Re:Show HN: Llama 3.3 70B Sparse Autoencoders with API access by trq_ on Monday December 23, 2024 @09:34PM (#42499262)
Hmm the hallucination would happen in the auto labelling, but we review and test our labels and they seem correct!