© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 15222
    Prev
    Next
    INDEX
    Explanations

    police and dynastic names

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

    The neuron strongly fires on capitalized names of ancient Near Eastern royals, deities, dynasties, and temple sites (e.g. Amon-Re, Hatshepsut, Karnak, Deir el-Bahri, Ibbi-Sin).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ?
    -1.71
    それにしても
    -1.54
     такой
    -1.45
    詳しくは
    -1.44
     现场
    -1.39
    z
    -1.39
     rumbling
    -1.38
    みたい
    -1.36
     is
    -1.34
    )
    -1.34
    POSITIVE LOGITS
    )=>{
    
    1.55
    ྜ
    1.53
    你
    1.50
    setMsg
    1.43
    鋮
    1.42
    ER
    1.40
     steadfast
    1.38
     even
    1.36
     freshwater
    1.34
    硖
    1.32
    Activations Density 0.018%

    No Known Activations