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

    Neuronpedia

    INDEX
    Explanations

    detailed explanations

    np_acts-logits-general · gemini-2.5-flash-lite

    phrases that signal formal, comprehensive explanation or in-depth, detailed exposition.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
    ziehungen
    0.61
    이가
    0.46
     встанов
    0.46
    ɐ
    0.46
     سوف
    0.45
    >("
    0.44
     вступи
    0.44
    Probit
    0.44
    \|=\
    0.44
    щихся
    0.44
    POSITIVE LOGITS
     battery
    0.41
     Battery
    0.41
     belts
    0.39
     hacks
    0.39
     Insulin
    0.39
     lens
    0.39
     Molecules
    0.38
     bands
    0.38
     on
    0.38
     earrings
    0.37
    Act Density 0.007%

    No Known Activations