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

    Neuronpedia

    INDEX
    Explanations

    passion, ionate, ion

    np_max-act · gemini-2.0-flash

    This neuron detects the “-ion” suffix, i.e. words ending in “ion” (as in “passion,” “implementation,” etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
     elt
    -0.07
     duck
    -0.07
     dikke
    -0.06
    �
    -0.06
    センター
    -0.06
     duplicates
    -0.06
    <dyn
    -0.06
    .getLocation
    -0.06
     учрежд
    -0.06
     unidentified
    -0.06
    POSITIVE LOGITS
     passion
    0.14
     Passion
    0.10
     passionate
    0.10
     passionately
    0.10
    asmine
    0.07
     aggression
    0.07
     passions
    0.07
    ZZ
    0.07
    900
    0.07
     Concepts
    0.07
    Act Density 0.010%

    No Known Activations