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

    Neuronpedia

    INDEX
    Explanations

    legitimate reasons/concerns/alternatives

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     étab
    1.85
     quedó
    1.84
    د
    1.81
    ين
    1.80
    ด์
    1.75
    이며
    1.69
    )}}
    1.67
    ując
    1.63
    رفته
    1.62
     mula
    1.57
    POSITIVE LOGITS
    ल
    1.92
    ли
    1.80
    veel
    1.77
     Saying
    1.63
    9
    1.63
    7
    1.62
    6
    1.60
    3
    1.59
    ik
    1.53
    ge
    1.53
    Act Density 0.003%

    No Known Activations