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

    Neuronpedia

    INDEX
    Explanations

    blog posts, articles

    np_max-act · gemini-2.0-flash

    The neuron predominantly activates on content‐bearing tokens—especially nouns, adjectives, and other informative words—rather than on common function words.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Negative Logits
    minecraft
    -0.06
    み
    -0.06
     oci
    -0.06
     owes
    -0.06
    داد
    -0.06
    िद
    -0.06
    ών
    -0.06
     گذاری
    -0.06
     kutje
    -0.06
    	onClick
    -0.06
    POSITIVE LOGITS
    stown
    0.07
     culprit
    0.07
     manuscripts
    0.06
     costly
    0.06
     Abort
    0.06
    bach
    0.06
    stretch
    0.06
    .Black
    0.06
     outlaw
    0.06
     StyleSheet
    0.06
    Act Density 0.186%

    No Known Activations