© 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. 54154
    Prev
    Next
    INDEX
    Explanations

    Research conducted

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

    The neuron strongly activates on programming‐language punctuation and operators (e.g. parentheses, braces, equals signs, semicolons).

    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
    ικοί
    -1.02
    ıyoruz
    -1.02
    ͻ
    -1.01
     vannak
    -1.00
     phénom
    -1.00
     euró
    -0.97
    información
    -0.96
     kabát
    -0.96
    Га
    -0.93
    Skal
    -0.93
    POSITIVE LOGITS
     here
    1.03
    <eos>
    0.98
     none
    0.94
     once
    0.90
     require
    0.88
     something
    0.87
     eens
    0.87
    forderungen
    0.85
     largely
    0.84
     must
    0.83
    Activations Density 0.017%

    No Known Activations