© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 9765
    Prev
    Next
    INDEX
    Explanations

    symbols or markers

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

    The neuron is keyed to isolated non-alphanumeric punctuation tokens (like `<` or `%`) that typically occur in code snippets—e.g. shell hints, placeholders, or regex constructs.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
    viš
    -1.56
    !(
    -1.44
    𖡼
    -1.39
    ⇹
    -1.38
    clientes
    -1.36
     if
    -1.34
    !』
    -1.34
    !!!!
    -1.32
    </tfoot>
    -1.31
     all
    -1.31
    POSITIVE LOGITS
     اين
    1.41
     beschränkt
    1.41
     vostre
    1.41
     واتس
    1.38
    بیش
    1.37
     festgelegt
    1.36
    茈
    1.35
    }\
    1.34
     ガード
    1.34
     Neutr
    1.34
    Activations Density 0.000%

    No Known Activations