© 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. 72300
    Prev
    Next
    INDEX
    Explanations

    compiling and running code

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

    The neuron is primarily triggered by numeric tokens (standalone numbers, numeric literals, and numeric flags).

    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
    entrum
    -1.06
     ayudar
    -1.00
    ֶּ
    -0.96
    Bains
    -0.96
    ִּ
    -0.94
    行动
    -0.90
    یب
    -0.88
    SuspendLayout
    -0.88
     brooms
    -0.88
     wrze
    -0.87
    POSITIVE LOGITS
     gigantes
    1.16
     poë
    1.16
    Източници
    1.13
     undervis
    1.11
     اين
    1.10
     potes
    1.09
     två
    1.04
    <>
    1.04
     stehe
    1.03
    AIRES
    1.02
    Activations Density 0.127%

    No Known Activations