© 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. 130277
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric tokens—especially version‐style numbers and other numeric literals—in code/import statements.

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

    test runner, extensions, mock

    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
     rozwoju
    -1.07
     realizado
    -1.07
     armados
    -1.04
     hjemmeside
    -1.03
     realizados
    -1.02
    ans
    -1.01
     preparado
    -1.00
     począ
    -0.99
     część
    -0.96
     lámpara
    -0.96
    POSITIVE LOGITS
    ktes
    1.07
    arashtra
    1.07
     fumig
    1.03
    ઓ
    1.02
     Орден
    0.99
    lunch
    0.98
    <<<<
    0.97
    
    0.97
    Ironically
    0.97
    Interestingly
    0.96
    Activations Density 0.001%

    No Known Activations