© 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. 22-GEMMASCOPE-RES-131K
    7. 104585
    Prev
    Next
    INDEX
    Explanations

    parts of objects, systems, and plants

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

    The neuron selectively activates on content-bearing nouns—especially concrete object or interface-element terms (like “background,” “column,” “heart,” “brackets,” “tab,” etc.).

    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
     insegu
    -0.98
     şeker
    -0.92
     fuß
    -0.90
     findViewById
    -0.88
    endence
    -0.85
     uzman
    -0.84
     infecciones
    -0.84
    的眼神
    -0.83
    education
    -0.83
    gebras
    -0.82
    POSITIVE LOGITS
    Gre
    0.94
    entuh
    0.88
     saluti
    0.86
     سات
    0.85
     ataupun
    0.84
     predomin
    0.84
     bibit
    0.83
    ureusement
    0.83
    )))),
    0.82
    局部
    0.81
    Activations Density 0.136%

    No Known Activations