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

    Neuronpedia

    INDEX
    Explanations

    phrases related to returning or regaining things or activities

    oai_token-act-pair · gpt-3.5-turbo

    references to returning or going back to a previous state or condition

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    uously
    -0.73
     understatement
    -0.71
    ashi
    -0.70
    uary
    -0.65
     Hots
    -0.62
     Guardiola
    -0.60
    cture
    -0.60
    "]=>
    -0.59
    ities
    -0.59
    idis
    -0.57
    POSITIVE LOGITS
    dated
    0.93
    doors
    0.89
    packs
    0.87
    GROUND
    0.87
    tracking
    0.84
    door
    0.83
    trace
    0.83
    stab
    0.82
    fired
    0.80
    haul
    0.78
    Act Density 0.030%

    No Known Activations