© 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. 39679
    Prev
    Next
    INDEX
    Explanations

    Vascular endothelial

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

    code structures involving loops and flow control like foreach, while, and echo statements.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron activates on standalone numeric literals (integers and floats) in code or text.

    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
     to
    -1.29
     of
    -1.29
    ших
    -1.00
     becomes
    -0.96
     gets
    -0.92
     does
    -0.92
     whose
    -0.91
    Frankreich
    -0.89
     has
    -0.89
    Արտաքին
    -0.89
    POSITIVE LOGITS
     then
    1.13
    就已经
    1.01
     пошли
    0.97
     immense
    0.96
    walking
    0.90
    either
    0.89
    ensed
    0.89
     seguido
    0.88
    鷥
    0.87
     현
    0.87
    Activations Density 0.001%

    No Known Activations