© 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. 6413
    Prev
    Next
    INDEX
    Explanations

    website headings

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron is specialized in detecting standalone numerical tokens (e.g. quantities, dates, measurements, and decimal numbers).

    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
    <start_of_image>
    0.61
    딥
    0.55
     NETWORK
    0.54
     PF
    0.54
    XXXXXXXX
    0.54
    Deep
    0.52
    RING
    0.52
     COVER
    0.52
    umap
    0.51
     Flats
    0.50
    POSITIVE LOGITS
    <h1>
    0.63
     }}$,
    0.56
    |+|
    0.55
    }|^{
    0.54
     আহ্বান
    0.52
    <h2>
    0.52
     }}^{\
    0.52
     चिंतन
    0.50
    ばん
    0.50
     category
    0.50
    Activations Density 0.036%

    No Known Activations