© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129613
    Prev
    Next
    INDEX
    Explanations

    The neuron reliably activates on numeric tokens or sequences of digits (e.g. version numbers, line numbers, file‐ID numbers, or image dimensions).

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

    filenames with numbers and extensions

    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
     what
    -1.45
     any
    -1.39
     their
    -1.33
     supremely
    -1.26
     create
    -1.25
     nestled
    -1.25
     get
    -1.23
     new
    -1.21
     or
    -1.20
     Get
    -1.19
    POSITIVE LOGITS
    ܆
    1.41
    鶿
    1.35
     ADDITIONAL
    1.33
    もちゃ
    1.31
    があるので
    1.31
     لهذه
    1.30
    
    1.27
    これは
    1.26
    eneral
    1.26
    この
    1.24
    Activations Density 0.011%

    No Known Activations