© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 1946
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on unusually long contiguous character sequences (long tokens or very long text spans).

    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
     backlinks
    0.51
    <unused67>
    0.47
    <unused50>
    0.46
    <unused92>
    0.46
    <unused82>
    0.46
    偈
    0.45
     경우에는
    0.45
    ዦ
    0.45
    দ্দিক
    0.44
    poetrylovers
    0.44
    POSITIVE LOGITS
    ;
    0.53
     +
    0.50
    +
    0.45
    Đặt
    0.43
     bunun
    0.40
    в
    0.39
    Phill
    0.39
    The
    0.38
    Term
    0.38
    putar
    0.38
    Activations Density 0.000%

    No Known Activations