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

    Neuronpedia

    INDEX
    Explanations

    phrases related to escaping dangers or threats

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    "Escape" or avoiding negative consequences

    np_acts-logits-general · gemini-2.0-flash

    escape from negative consequences

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    :][
    -0.65
     BoxFit
    -0.64
    NgModule
    -0.60
    TagMode
    -0.59
    Datuak
    -0.59
    AlterField
    -0.59
    __':
    -0.57
    "]));
    -0.57
    createComponent
    -0.56
    yship
    -0.56
    POSITIVE LOGITS
     attacks
    0.94
     scrutiny
    0.91
     attack
    0.81
     onslaught
    0.78
    attacks
    0.75
     pressures
    0.68
    attack
    0.67
     threats
    0.67
     criticism
    0.66
     pressure
    0.66
    Act Density 0.484%

    No Known Activations