INDEX
    Explanations

    abbreviations

    The neuron selectively activates on fragments of proper names or acronyms (i.e. tokens belonging to capitalized terms such as “Específicos,” “Arachis hypogaea,” “Indicator,” etc.).

    New Auto-Interp
    Negative Logits
    .marker
    -0.06
     AssemblyTitle
    -0.06
    instanc
    -0.06
    เตร
    -0.06
    ();↵↵↵↵
    -0.06
    .writeObject
    -0.06
     minimalist
    -0.06
     }?>↵
    -0.06
    -describedby
    -0.06
    Hibernate
    -0.06
    POSITIVE LOGITS
     Microwave
    0.07
     crop
    0.07
     přiroz
    0.06
    (destination
    0.06
    циональ
    0.06
    _visible
    0.06
    yat
    0.06
    0.06
    ريط
    0.06
    mins
    0.06
    Act Density 0.103%

    No Known Activations