© 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. 10-GEMMASCOPE-RES-131K
    7. 105619
    Prev
    Next
    INDEX
    Explanations

    programming operators and keywords

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

    the ampersand symbol (&) in code, particularly in contexts involving references, pointers, or type constraints.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron activates on ampersand characters (“&”) in code, highlighting reference patterns, pointer types, or generic-constraint symbols.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     if
    -1.62
    {
    -1.58
     ‘
    -1.53
    k
    -1.52
    几分
    -1.51
    is
    -1.48
    鸰
    -1.39
     '
    -1.38
     
    -1.34
     robić
    -1.32
    POSITIVE LOGITS
     of
    1.69
     herren
    1.53
     possiamo
    1.52
     Politica
    1.48
    kadang
    1.48
     kutu
    1.48
     zustand
    1.48
    interopRequire
    1.47
     oock
    1.46
     kijang
    1.46
    Activations Density 0.001%

    No Known Activations