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

    Neuronpedia

    INDEX
    Explanations

    @

    np_max-act · gemini-2.0-flash

    The neuron is spotting email‐header tokens (addresses and routing markers, e.g. “@”, usernames, “Forwarded by”, date/time lines, etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     SCRIPT
    -0.07
     Bed
    -0.07
    CON
    -0.07
     Commit
    -0.07
     Types
    -0.07
     CON
    -0.06
     File
    -0.06
     blueprint
    -0.06
    Ign
    -0.06
    	board
    -0.06
    POSITIVE LOGITS
     jpeg
    0.07
    éf
    0.07
     separating
    0.06
    ！');↵
    0.06
    ीट
    0.06
     below
    0.06
     있다
    0.06
    })}↵
    0.05
    azes
    0.05
    (disposing
    0.05
    Act Density 0.003%

    No Known Activations