© 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. 98457
    Prev
    Next
    INDEX
    Explanations

    code and natural language

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

    The neuron activates on numeric literals (digits) in code, especially when they appear in input prompts or as numeric constants.

    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
    残
    -0.91
    NoSuchAlgorithm
    -0.89
    вів
    -0.86
    燃
    -0.84
    見えます
    -0.84
     morgon
    -0.83
     africa
    -0.83
    Sucesor
    -0.83
     only
    -0.82
     each
    -0.82
    POSITIVE LOGITS
     if
    1.04
     если
    1.01
     inserire
    0.98
     suministro
    0.97
     WHILE
    0.95
    Frankfurt
    0.95
     אם
    0.92
    öten
    0.92
     enthou
    0.90
     avete
    0.90
    Activations Density 0.003%

    No Known Activations