© 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. 34-GEMMASCOPE-RES-131K
    7. 31978
    Prev
    Next
    INDEX
    Explanations

    approximating or approximating concepts

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

    This neuron fires on the technical‐problem headings—labels or titles of math/programming questions (e.g. “Numerical iterative method…,” “Error on first run…,” “Correctness of the integral…,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     GMAT
    -0.80
    確認ください
    -0.73
    حة
    -0.73
    getDriver
    -0.72
     Geschäfte
    -0.72
     frecuentemente
    -0.72
    创始人
    -0.71
     Preise
    -0.71
    tetur
    -0.69
     Okt
    -0.69
    POSITIVE LOGITS
     Approximation
    1.05
     approximations
    0.96
     approximation
    0.94
     aproxima
    0.92
     approximate
    0.86
     approximates
    0.85
     central
    0.85
    xl
    0.85
     table
    0.85
    컴
    0.82
    Activations Density 0.022%

    No Known Activations