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

    Neuronpedia

    INDEX
    Explanations

    intentionally bad reward model

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     верну
    0.48
     තු
    0.48
    결
    0.45
     পারিল
    0.45
     ಮತ್ತೆ
    0.45
     یک
    0.45
     পুনরায়
    0.44
     بین
    0.44
    лизова
    0.44
    ವಹ
    0.44
    POSITIVE LOGITS
    uk
    0.48
    дзен
    0.44
    以外
    0.42
    present
    0.42
     Lack
    0.42
    prefer
    0.41
    mazing
    0.41
    缺乏
    0.40
    ac
    0.40
     Apart
    0.40
    Act Density 0.001%

    No Known Activations