Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    As such refusal

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Optionally
    0.44
     quoique
    0.39
     থাকিলেও
    0.37
    WELL
    0.37
     wel
    0.37
     ",")
    0.37
     Though
    0.36
    আবার
    0.36
    Well
    0.36
     well
    0.36
    POSITIVE LOGITS
     such
    0.87
    such
    0.68
    sembled
    0.65
     SUCH
    0.60
    sembling
    0.59
     Such
    0.53
     такие
    0.52
    bestos
    0.52
     part
    0.51
    这样的
    0.51
    Act Density 0.017%

    No Known Activations