© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5921
    Prev
    Next
    INDEX
    Explanations

    Dota 2 and passwords

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

    The neuron fires strongly on numeric address components (house numbers, street names/avenues, ZIP‐like digit sequences)—i.e. it detects street addresses.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Anambra
    0.93
     Mauritius
    0.82
    Rajasthan
    0.77
    meida
    0.73
    🇱
    0.72
    Sao
    0.70
     sao
    0.70
    睢
    0.69
     portug
    0.68
    🐊
    0.67
    POSITIVE LOGITS
     Seattle
    2.80
    Seattle
    2.63
     Puget
    2.45
     Tacoma
    2.28
     Spokane
    2.23
     Vancouver
    1.91
     Yakima
    1.81
     Oregon
    1.77
    Vancouver
    1.77
     Bellingham
    1.73
    Activations Density 0.407%

    No Known Activations