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

    Neuronpedia

    INDEX
    Explanations

    the use of the word "perhaps" and its variations in various contexts, indicating uncertainty or speculation

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

    Preceding "perhaps" (case-insensitive)

    np_acts-logits-general · gemini-2.0-flash

    introducing a speculation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     apparently
    -0.83
     evidently
    -0.75
     reportedly
    -0.74
     obviously
    -0.74
    apparently
    -0.72
     presumably
    -0.72
     undoubtedly
    -0.72
     Apparently
    -0.69
     ovviamente
    -0.68
    presumably
    -0.68
    POSITIVE LOGITS
    anskje
    1.11
     someday
    0.93
    المناصب
    0.86
    ValueStyle
    0.83
     jopa
    0.83
    GEBURTSDATUM
    0.82
     even
    0.80
     sogar
    0.76
     propOrder
    0.76
     insuffisamment
    0.73
    Act Density 0.243%

    No Known Activations