© 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. 13153
    Prev
    Next
    INDEX
    Explanations

    long string or padding

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

    The neuron fires on short duration‐related phrases, especially “How long…” questions or “not available” statements.

    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
     lø
    0.68
     Tôi
    0.63
     Ingles
    0.61
     レディース
    0.61
     ૧
    0.60
     Règles
    0.59
     Jeu
    0.59
     আর্জেন্টিনা
    0.58
     Đi
    0.58
     النظام
    0.58
    POSITIVE LOGITS
    language
    0.66
    history
    0.57
    Historic
    0.57
    erms
    0.57
    텁
    0.57
    transmission
    0.56
    getSave
    0.55
    kHz
    0.55
    harp
    0.55
    survey
    0.54
    Activations Density 0.000%

    No Known Activations