© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 26800
    Prev
    Next
    INDEX
    Explanations

    buffer, array, matrix, data

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

    The neuron fires on C++-style type declarations, especially template-based types with angle brackets and reference symbols (e.g. “std::vector<uint8_t>& buf”).

    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
     first
    -1.67
     after
    -1.63
     a
    -1.58
     two
    -1.49
     on
    -1.42
     three
    -1.41
     answer
    -1.41
     giving
    -1.41
     “
    -1.35
     give
    -1.35
    POSITIVE LOGITS
     barbacoa
    1.61
     berge
    1.60
     Kopi
    1.50
     Ayam
    1.46
     bandar
    1.42
     Situs
    1.41
     thermomix
    1.41
     Jasa
    1.41
     uchiha
    1.40
     boute
    1.39
    Activations Density 0.023%

    No Known Activations