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

    Neuronpedia

    INDEX
    Explanations

    phrases that indicate concessive or contrasting information

    oai_token-act-pair · gpt-3.5-turbo

    phrases that include the word "admittedly" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    lets
    -0.80
    ogy
    -0.73
    seller
    -0.71
    eding
    -0.67
    arij
    -0.66
    let
    -0.65
    tein
    -0.64
    eng
    -0.63
    irled
    -0.63
    lines
    -0.63
    POSITIVE LOGITS
     admittedly
    0.79
    âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ
    0.79
    odox
    0.75
    imaru
    0.75
    mittedly
    0.74
    âĶĢâĶĢâĶĢâĶĢ
    0.72
    entimes
    0.70
    itably
    0.69
    ãĤº
    0.68
    ãĥ¼ãĥĨ
    0.67
    Act Density 0.005%

    No Known Activations