INDEX
    Explanations

    high importance

    This neuron responds to evaluative and intensifying words—adjectives and adverbs that mark emphasis or promotion (e.g. “most,” “promising,” “greatly,” “urgently”).

    New Auto-Interp
    Negative Logits
     giờ
    -0.08
     wo
    -0.07
                      
    -0.07
    -0.07
                       
    -0.06
    Bit
    -0.06
     dictates
    -0.06
    endregion
    -0.06
    -0.06
              
    -0.06
    POSITIVE LOGITS
     갤로그
    0.06
     offsetX
    0.06
     nutné
    0.06
     ''}↵
    0.06
    nova
    0.06
     ))}↵
    0.06
     das
    0.06
    )>↵
    0.06
     이동
    0.06
     Nová
    0.06
    Act Density 0.059%

    No Known Activations