© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 39485
    Prev
    Next
    INDEX
    Explanations

    php, french, spanish, german, italian, turkish

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

    This neuron responds to code‐style or symbol‐laden tokens—strings that mix letters with punctuation (e.g. filenames like “retrieve.php,” quantum‐state notations like “|0〉s,” hyphenated or dotted identifiers).

    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
     increases
    -0.92
     every
    -0.85
    cortina
    -0.84
     out
    -0.82
     for
    -0.82
     any
    -0.81
    になりたい
    -0.81
    韌
    -0.80
     under
    -0.79
     taking
    -0.79
    POSITIVE LOGITS
     öf
    0.92
    ینگ
    0.89
    ğraf
    0.88
    mdash
    0.88
     distinguer
    0.88
     construire
    0.88
    catcher
    0.88
     beso
    0.87
     Padova
    0.87
     maler
    0.87
    Activations Density 0.018%

    No Known Activations