© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 9622
    Prev
    Next
    INDEX
    Explanations

    typescript code blocks

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

    The neuron strongly responds to short, punctuation-free alphabetic tokens—especially acronyms, author initials, or file-extension markers (e.g. “Witt,” “WIT,” “js”).

    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
    టే
    0.52
    astava
    0.51
    ೕ
    0.50
     Rakh
    0.50
    ட்டான
    0.50
     ih
    0.49
     angezeigt
    0.49
     cilantro
    0.49
    iscale
    0.49
    冭
    0.49
    POSITIVE LOGITS
    //#
    0.57
     কমিশন
    0.55
    汽车
    0.55
    appointments
    0.54
    समा
    0.52
    leads
    0.52
     সৈন্য
    0.51
    blasts
    0.51
    comple
    0.50
    汽車
    0.50
    Activations Density 0.010%

    No Known Activations