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

    Neuronpedia

    INDEX
    Explanations

    The neuron seems to be looking for references to specific entities or attributes indicated by the word "that" with an emphasis on explanations or relationships

    oai_token-act-pair · gpt-3.5-turbo

    the word "that" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    brates
    -0.76
    ormons
    -0.76
    rior
    -0.70
    cycles
    -0.69
    uously
    -0.69
    oby
    -0.69
    istics
    -0.69
    ciples
    -0.68
    asters
    -0.68
     Leilan
    -0.66
    POSITIVE LOGITS
     pesky
    1.12
     fateful
    1.08
     particular
    1.04
     same
    0.97
     kind
    0.97
     sort
    0.85
     equation
    0.84
     aforementioned
    0.84
     elusive
    0.83
     type
    0.81
    Act Density 0.102%

    No Known Activations