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

    Neuronpedia

    INDEX
    Explanations

    references to locations such as "basement."

    oai_token-act-pair · gpt-3.5-turbo

    repeated mentions of "basement."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    nir
    -0.77
    uncture
    -0.74
    Cola
    -0.71
    acting
    -0.70
    utan
    -0.67
    uality
    -0.67
    weather
    -0.66
    drawn
    -0.66
    inel
    -0.66
    Forward
    -0.65
    POSITIVE LOGITS
     basement
    1.18
     cellar
    0.97
    stairs
    0.96
     annex
    0.89
     Dwell
    0.88
    ħĭ
    0.86
     lid
    0.85
     loft
    0.84
     attic
    0.84
     courtyard
    0.83
    Act Density 0.009%

    No Known Activations