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

    Neuronpedia

    INDEX
    Explanations

    attends to specific parameter values from contextual tokens within a technical or instruction-based framework

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.07
    1:0.09
    2:0.07
    3:0.07
    4:0.10
    5:0.04
    6:0.19
    7:0.33
    Negative Logits
     littéraire
    -0.31
     sangue
    -0.27
    acetic
    -0.26
     الوطنيه
    -0.25
    näm
    -0.24
     beszél
    -0.24
     tilgjenge
    -0.24
     nucléaire
    -0.23
     batteria
    -0.23
     ácidos
    -0.23
    POSITIVE LOGITS
     */;
    0.48
    parsedMessage
    0.46
     }}"></
    0.45
    Parcelize
    0.44
     سكانية
    0.44
    MLLoader
    0.44
    __;
    0.43
    /*
    0.43
     AssemblyCompany
    0.42
     invokingState
    0.42
    Act Density 0.657%

    No Known Activations