공대생 도전 일지

SAE를 통해 LLM의 데이터를 변경, 조작해보자가 시작되었습니다!!!!https://transformer-circuits.pub/2024/scaling-monosemanticity/ Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 SonnetAuthors Adly Templeton*, Tom Conerly*, Jonathan Marcus, Jack Lindsey, Trenton Bricken, Brian Chen, Adam Pearce, Craig Citro, Emmanuel Ameisen, Andy Jones, Hoagy Cunningham, Nicholas L Turner, Callum McDougall, Monte ..

https://transformer-circuits.pub/2022/toy_model/ Toy Models of SuperpositionIt would be very convenient if the individual neurons of artificial neural networks corresponded to cleanly interpretable features of the input. For example, in an “ideal” ImageNet classifier, each neuron would fire only in the presence of a specific vtransformer-circuits.pub 제가 원했던 분야 중 하나입니다.드디어 조금씩 진행하게 되었네요 일부 뉴런은 명확..