© 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. 37953
    Prev
    Next
    INDEX
    Explanations

    US state abbreviations

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

    The neuron fires on place names—particularly U.S. city names (often in legal addresses) and their accompanying state abbreviations.

    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
     Reverend
    -0.84
    ه‌ی
    -0.79
     सो
    -0.75
     $(
    -0.75
    š
    -0.73
     Figure
    -0.72
    議院
    -0.71
    \_
    -0.70
     peringatan
    -0.69
     Professor
    -0.69
    POSITIVE LOGITS
     Assn
    1.28
     Calif
    1.25
     Eqs
    1.20
     Vt
    1.02
     Univ
    1.02
     Md
    0.99
     Okla
    0.99
    Calif
    0.97
     alemão
    0.96
     Assoc
    0.95
    Activations Density 0.099%

    No Known Activations