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

    Neuronpedia

    INDEX
    Explanations

    base

    np_max-act · gemini-2.0-flash

    The neuron responds to uncommon, domain‐specific or technical words (i.e. rarer vocabulary) rather than everyday terms.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    ケ
    -0.07
     kontakt
    -0.06
    /sys
    -0.06
    OutOfBoundsException
    -0.06
    Tell
    -0.06
     surv
    -0.06
     turn
    -0.06
    .Entity
    -0.06
     件
    -0.06
     Shame
    -0.06
    POSITIVE LOGITS
     Exercises
    0.07
     antennas
    0.07
     emergencies
    0.06
     ebony
    0.06
     grads
    0.06
    VICES
    0.06
    ENCIL
    0.06
    trag
    0.06
     magn
    0.06
    chlor
    0.06
    Act Density 0.676%

    No Known Activations