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

    Neuronpedia

    INDEX
    Explanations

    programmed refusal of requests

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    末
    0.59
    με
    0.55
    িত্য
    0.54
    或者
    0.53
    即便
    0.52
    DVRIP
    0.52
    Despatx
    0.50
    集
    0.50
    其
    0.50
    특
    0.48
    POSITIVE LOGITS
     sorry
    0.79
     glad
    0.79
     afraid
    0.69
     grateful
    0.69
     proud
    0.64
     heartbroken
    0.64
     not
    0.63
     am
    0.63
     trying
    0.63
     akan
    0.63
    Act Density 0.136%

    No Known Activations