© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128630
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on tokens that introduce or frame quantitative or measured statements (e.g. distances, rates, problems arising, specifications).

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

    foreign language or english words

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     of
    -1.34
     was
    -1.13
     lately
    -1.11
     although
    -1.09
    もしか
    -1.08
     currently
    -1.02
    zką
    -1.02
     strukt
    -1.01
    <bos>
    -0.99
     even
    -0.99
    POSITIVE LOGITS
    様な
    1.25
     görmek
    1.18
     évit
    1.16
     Männern
    1.13
    ウィキ
    1.13
    臈
    1.10
     laissant
    1.08
     permitiendo
    1.06
     Now
    1.05
    这么多
    1.05
    Activations Density 0.002%

    No Known Activations