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

    Neuronpedia

    INDEX
    Explanations

    pointer

    np_max-act · gemini-2.0-flash

    This neuron detects C/C++ pointer‐and‐address operations, especially C‐style casts that mix “&”, “*”, and a type in parentheses (e.g. `(float*)&…` or `*(float*)&…`).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    นอ
    -0.06
    loader
    -0.06
    。しかし
    -0.06
     πριν
    -0.06
     Macron
    -0.06
    capitalize
    -0.06
     comprar
    -0.06
     πά
    -0.06
    стер
    -0.06
     turret
    -0.05
    POSITIVE LOGITS
     hierarchy
    0.08
    .getMessage
    0.07
     lawsuits
    0.07
     Conn
    0.07
    �
    0.07
    ierarchy
    0.06
    \F
    0.06
    //------------------------------------------------------------------------------↵
    0.06
     WB
    0.06
    efined
    0.06
    Act Density 0.024%

    No Known Activations