© 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. 13616
    Prev
    Next
    INDEX
    Explanations

    concepts of state and nature

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

    The neuron strongly activates on multi‐syllabic, content-bearing nouns—especially technical, product, or domain-specific terms.

    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
    在我
    0.66
    𝒇
    0.65
     说
    0.62
    ｱ
    0.62
     俺
    0.61
    <0x0F>
    0.60
     Conan
    0.59
    Ҡ
    0.59
    <html>
    0.58
     Honestly
    0.58
    POSITIVE LOGITS
     نہیں۔
    0.91
    }.
    0.86
     ہو۔
    0.84
    ।
    0.83
    是因为
    0.80
    是因為
    0.80
    ’।
    0.80
    ”।
    0.79
    」。
    0.79
    ++.
    0.77
    Activations Density 0.329%

    No Known Activations