© 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. 34-GEMMASCOPE-RES-131K
    7. 57258
    Prev
    Next
    INDEX
    Explanations

    clerk

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

    The neuron specializes in spotting document header metadata lines—those long runs of uppercase tokens (e.g. “FILED,” dates, “UNITED STATES COURT OF APPEALS,” file numbers, clerk names) typical of legal filing headers.

    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
     指
    -0.75
    mcn
    -0.74
    rehen
    -0.72
    élien
    -0.72
    となりました
    -0.71
     memerlukan
    -0.69
    vinner
    -0.68
    Sometimes
    -0.68
    komna
    -0.68
    今でも
    -0.67
    POSITIVE LOGITS
     clerk
    4.09
     Clerk
    3.89
    Clerk
    3.27
     clerks
    2.89
    cler
    2.48
     CLERK
    2.42
     Clerks
    2.39
     secretary
    2.23
     Secretary
    1.97
     cler
    1.92
    Activations Density 0.229%

    No Known Activations