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

    Neuronpedia

    INDEX
    Explanations

    GLUE benchmark

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     multipliers
    0.40
     touches
    0.39
     touched
    0.38
     touching
    0.37
    CHARS
    0.36
     prays
    0.36
     restarted
    0.35
     zb
    0.35
     troubles
    0.35
     mechanized
    0.35
    POSITIVE LOGITS
    shore
    0.42
    <0xC4>
    0.41
    冰箱
    0.41
    etal
    0.41
     iceberg
    0.41
     ghat
    0.41
    銀行
    0.40
    beige
    0.40
    ูล
    0.40
    доо
    0.40
    Act Density 0.003%

    No Known Activations