© 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. 5366
    Prev
    Next
    INDEX
    Explanations

    cuda device

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

    The neuron strongly activates on quotation‐mark tokens (i.e. opening or closing quote characters) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    opening quotation marks that begin direct quotes or strings in code and text.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    opening quotation marks that introduce quoted text or string literals.

    oai_token-act-pair · gpt-5Triggered 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
     “…
    0.89
     Kidney
    0.86
     Conversely
    0.85
     Continuando
    0.85
     えっと
    0.84
     Kemudian
    0.84
     Renal
    0.83
     或者
    0.83
     LANGUAGES
    0.83
     Apartments
    0.83
    POSITIVE LOGITS
    https
    1.15
    http
    1.06
    ка
    0.95
    КА
    0.88
    Пре
    0.86
    ба
    0.85
    Вы
    0.83
    Ро
    0.83
    в
    0.82
    none
    0.82
    Activations Density 0.011%

    No Known Activations