© 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. 33059
    Prev
    Next
    INDEX
    Explanations

    David Echevaria

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

    The neuron is strongly activated by numeric tokens (digits and years), so it detects numbers in the text.

    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
    ToBe
    -1.02
     biß
    -0.96
    pyp
    -0.96
     Teraz
    -0.94
    attuale
    -0.92
     yanlış
    -0.91
    瞩
    -0.90
     حتی
    -0.90
    ڤ
    -0.90
    sApp
    -0.89
    POSITIVE LOGITS
     didn
    1.02
    ϳ
    1.02
     don
    1.01
     روان
    1.00
     salaries
    0.97
    leich
    0.96
     diga
    0.96
     Wednesday
    0.95
     county
    0.95
    inosau
    0.94
    Activations Density 0.021%

    No Known Activations