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

    Neuronpedia

    INDEX
    Explanations

    words related to rules, documentation, and communication in formal settings

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Negative Logits
     mosqu
    -0.84
     stricken
    -0.80
     guiActiveUnfocused
    -0.74
     descending
    -0.74
     condol
    -0.72
     Danish
    -0.72
     detached
    -0.71
     Golem
    -0.70
     harmless
    -0.70
     nearest
    -0.70
    POSITIVE LOGITS
    ¹
    1.00
    £
    0.98
    âĹ
    0.94
    º
    0.93
    tm
    0.92
    »
    0.92
    ¡
    0.91
    hs
    0.90
    ¯¯¯¯
    0.89
    ®
    0.87
    Act Density 7.650%

    No Known Activations