© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 54211
    Prev
    Next
    INDEX
    Explanations

    place names and judges

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

    The neuron fires on the all‐caps place names and official titles that form a news article’s dateline header.

    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
     nessun
    -0.77
     Theodor
    -0.73
    чную
    -0.72
    rolla
    -0.72
     gérer
    -0.71
     తె
    -0.69
     nessuna
    -0.69
    "}"
    -0.69
     milky
    -0.68
     Shrine
    -0.68
    POSITIVE LOGITS
     Binghamton
    1.15
     Rockford
    0.89
    Riverside
    0.87
     Broome
    0.82
     Riverside
    0.81
     Redding
    0.80
     Greenville
    0.80
     fluoride
    0.79
    quehanna
    0.77
     Furman
    0.77
    Activations Density 0.015%

    No Known Activations