© 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. 128392
    Prev
    Next
    INDEX
    Explanations

    This neuron specializes in spotting structural or formatting tokens (e.g. XML/HTML tags, markup delimiters, code-like tokens) rather than ordinary text words.

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

    tokens 4 Controller POST INSTALL

    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
    に行って
    -0.96
     powerful
    -0.86
     civil
    -0.83
    omenclature
    -0.81
     that
    -0.81
     wider
    -0.78
     fulfilling
    -0.78
    に行きました
    -0.77
    AutoComplete
    -0.77
     veröffentlicht
    -0.76
    POSITIVE LOGITS
     بتوان
    0.84
     Sakit
    0.83
     anthology
    0.82
    周泽楷
    0.80
     entro
    0.77
    バージョン
    0.77
     पढ़
    0.77
     Huerta
    0.76
     académico
    0.76
    temple
    0.76
    Activations Density 0.016%

    No Known Activations