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

    Neuronpedia

    INDEX
    Explanations

    family members and gendered nouns

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     dunno
    0.43
    ექს
    0.42
     alakalı
    0.41
     koş
    0.40
    に応じて
    0.40
     downwards
    0.40
    imgur
    0.40
    を使う
    0.40
     dudes
    0.40
    imbangkan
    0.40
    POSITIVE LOGITS
     preceded
    0.84
     predeceased
    0.82
     beloved
    0.71
     cherished
    0.68
     loving
    0.67
     lovingly
    0.66
     Survivors
    0.65
     survives
    0.63
     resided
    0.62
     survived
    0.62
    Act Density 0.004%

    No Known Activations