© 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. 34-GEMMASCOPE-RES-131K
    4. 107437
    Prev
    Next
    INDEX
    Explanations

    writing followed by a comma

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

    The neuron is highly tuned to the word “writing,” especially in the legal phrase “agreed to in writing.”

    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
    Eastern
    -0.73
    Xc
    -0.72
     labs
    -0.67
    笛
    -0.67
     Biblia
    -0.66
    次郎
    -0.66
    zuje
    -0.65
    خاب
    -0.65
     torneo
    -0.65
     McCall
    -0.65
    POSITIVE LOGITS
     Solis
    0.81
     marlin
    0.68
     tabby
    0.67
     Pry
    0.67
     lf
    0.67
    rière
    0.66
     keuken
    0.66
     inductor
    0.66
     중
    0.66
     culvert
    0.66
    Activations Density 0.062%

    No Known Activations