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

    Neuronpedia

    INDEX
    Explanations

    the word "taken" and its variations, indicating references to actions or situations involving something being taken or removed

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     gave
    -1.14
     saw
    -1.01
     took
    -1.00
     did
    -0.94
     was
    -0.87
     threw
    -0.77
    gave
    -0.76
     went
    -0.75
     came
    -0.74
     showed
    -0.71
    POSITIVE LOGITS
     taken
    1.63
    taken
    1.60
     flown
    1.59
     gone
    1.58
     Taken
    1.54
     arisen
    1.51
    Taken
    1.48
    gone
    1.47
     spoken
    1.46
    Seen
    1.45
    Act Density 0.211%

    No Known Activations