© 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. 22-GEMMASCOPE-RES-131K
    4. 104468
    Prev
    Next
    INDEX
    Explanations

    locations

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

    The neuron flags auxiliary verbs (especially in passive‐voice constructions) that commonly appear in formal announcements—e.g. “will,” “be,” “held,” “was,” “were,” “been.”

    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
    าหลี
    -0.90
     там
    -0.89
     Fucking
    -0.88
    Здесь
    -0.86
    Там
    -0.85
     Colonies
    -0.85
    ┍
    -0.84
     Shea
    -0.82
    वन
    -0.82
    acão
    -0.81
    POSITIVE LOGITS
     located
    1.84
     locations
    1.51
    located
    1.49
    場所は
    1.45
     location
    1.37
    Located
    1.36
     Location
    1.35
     Locations
    1.34
     Located
    1.32
    Location
    1.27
    Activations Density 0.272%

    No Known Activations