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

    Neuronpedia

    INDEX
    Explanations

    mandatory, essential, crucial

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     deserves
    1.12
     relates
    0.94
     reasonableness
    0.91
     eignet
    0.90
     rational
    0.89
     aligns
    0.88
     deserve
    0.87
     lends
    0.86
     contributes
    0.85
     aligning
    0.85
    POSITIVE LOGITS
     Mandatory
    1.04
    Mand
    1.01
     MAND
    0.92
     obligatoire
    0.91
     mandatory
    0.89
    强制
    0.79
     अनिवार्य
    0.78
    mandatory
    0.77
     obligatorio
    0.76
     oblig
    0.74
    Act Density 0.194%

    No Known Activations