© 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. 24-GEMMASCOPE-2-RES-16K
    7. 1981
    Prev
    Next
    INDEX
    Explanations

    Shadowberry Preserve; key values mismatch; feeling lost

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

    The neuron is specialized to detect tokens that are numeric or contain digits.

    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_24_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
    <unused2206>
    0.51
     assailant
    0.50
    lur
    0.48
     massacre
    0.45
     và
    0.45
     və
    0.45
     Kapol
    0.45
     и
    0.44
     नतीजे
    0.44
     urn
    0.43
    POSITIVE LOGITS
     부터
    0.48
     puente
    0.47
     পরবর্ত
    0.45
     pensamento
    0.44
    ண்மை
    0.44
    人們
    0.44
     الانسان
    0.44
     仕事
    0.43
    🧶
    0.43
    <html>
    0.43
    Activations Density 0.002%

    No Known Activations