© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 13731
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on URL‐style tokens—especially forward slashes following domain names—so it’s detecting web‐address patterns.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     desg
    0.62
     Herzegovina
    0.57
     pavil
    0.55
     "":
    0.54
     الأن
    0.54
    瑟
    0.54
     categorize
    0.53
     কান্না
    0.53
    ٴ
    0.53
    த்தல்
    0.52
    POSITIVE LOGITS
    events
    0.58
     события
    0.58
    Routes
    0.56
    GRA
    0.55
    Events
    0.55
    5
    0.55
    Space
    0.53
     divalent
    0.53
    iei
    0.52
    for
    0.52
    Activations Density 0.000%

    No Known Activations