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

    Neuronpedia

    INDEX
    Explanations

    part

    np_max-act · gemini-2.0-flash

    This neuron detects hedging or qualifying phrases (e.g. “may,” “at,” “least,” “in,” “part”) used to qualify scientific claims.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    Pocket
    -0.07
     Kurt
    -0.07
     poměr
    -0.07
    hcp
    -0.06
    лося
    -0.06
     Rat
    -0.06
    хов
    -0.06
     hass
    -0.06
    Notification
    -0.06
    Func
    -0.06
    POSITIVE LOGITS
     boş
    0.07
     اولین
    0.06
     examines
    0.06
    _Instance
    0.06
    　　　　　　
    0.06
    Replace
    0.06
    	video
    0.06
    Visit
    0.06
     recovered
    0.06
    assist
    0.06
    Act Density 0.012%

    No Known Activations