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

    Neuronpedia

    INDEX
    Explanations

    fragment

    np_max-act · gemini-2.0-flash

    This neuron detects occurrences of the word “fragment” (in any form or morphological variant, e.g., Fragment, fragments, fragmentation).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    initely
    -0.07
    ouis
    -0.07
    (on
    -0.07
    ponential
    -0.07
     Lee
    -0.07
    こそ
    -0.07
    oi
    -0.06
     Louise
    -0.06
     Cincinnati
    -0.06
     Russell
    -0.06
    POSITIVE LOGITS
     fragments
    0.10
    frag
    0.10
     Frag
    0.09
     fragment
    0.09
     frag
    0.09
     fragmentation
    0.08
     fragile
    0.08
     fragmented
    0.08
     Fragment
    0.08
    /fr
    0.08
    Act Density 0.009%

    No Known Activations