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

    Neuronpedia

    INDEX
    Explanations

    phrases related to unexpected negative consequences or outcomes

    oai_token-act-pair · gpt-3.5-turbo

    terms related to failures or adverse outcomes, particularly involving "fire" as a metaphor

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
    member
    -0.70
    agar
    -0.69
    said
    -0.68
    hi
    -0.68
    anson
    -0.65
    uala
    -0.64
    sal
    -0.63
    nih
    -0.63
    uv
    -0.63
    wra
    -0.62
    POSITIVE LOGITS
     tactics
    0.68
    aneously
    0.68
    iveness
    0.66
     Healer
    0.66
     Moreno
    0.65
    姫
    0.65
    ANCE
    0.65
    icult
    0.63
    ļéĨĴ
    0.63
    acies
    0.62
    Act Density 0.097%

    No Known Activations