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

    Neuronpedia

    INDEX
    Explanations

    attends to the latest closing double asterisk token from a series of tokens that appear later in the sentence

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Head Attr Weights
    0:0.04
    1:0.06
    2:0.07
    3:0.55
    4:0.07
    5:0.07
    6:0.06
    7:0.04
    Negative Logits
    MLLoader
    -0.40
     Италијани
    -0.37
     disambiguazione
    -0.36
    Rohy
    -0.35
     otomatig
    -0.33
     "..\..\..\
    -0.33
    Autoritní
    -0.32
    حياتها
    -0.32
     Earle
    -0.31
     урна
    -0.31
    POSITIVE LOGITS
     CURIAM
    0.43
     Rptr
    0.37
    umptive
    0.37
    verwijspagina
    0.36
     Imo
    0.33
    ggiare
    0.32
    Xie
    0.32
    /*---
    0.31
    RegressionTest
    0.31
    viedo
    0.30
    Act Density 0.012%

    No Known Activations