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

    Neuronpedia

    INDEX
    Explanations

    expressions and phrases indicating requests or responses

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.02
    1:0.02
    2:0.04
    3:0.07
    4:0.15
    5:0.03
    6:0.05
    7:0.34
    8:0.03
    9:0.04
    10:0.07
    11:0.08
    Negative Logits
     trenches
    -1.76
    ypes
    -1.55
    ��
    -1.55
    vantage
    -1.55
    igree
    -1.52
    ocamp
    -1.51
    vironments
    -1.50
     groove
    -1.49
    kered
    -1.42
    ourney
    -1.42
    POSITIVE LOGITS
     inquiries
    1.92
     allegations
    1.78
     queries
    1.73
     criticism
    1.71
     requests
    1.66
     inaction
    1.66
     petitions
    1.61
     inquiry
    1.58
     baseless
    1.57
     criticisms
    1.55
    Act Density 0.009%

    No Known Activations