© 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. 34-GEMMASCOPE-RES-131K
    4. 107817
    Prev
    Next
    INDEX
    Explanations

    beginnings and origins

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

    This neuron fires on build‐system or code‐repository lines—specifically variable‐assignment lines (like “mpi_sources += …”) and their continuations in makefiles or patch diffs.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    azón
    -0.80
     phải
    -0.78
    frac
    -0.78
     vastaan
    -0.75
    ровке
    -0.75
    Vtbl
    -0.74
    ším
    -0.72
     døde
    -0.72
    mohon
    -0.71
    allee
    -0.71
    POSITIVE LOGITS
    $\
    2.58
     $\
    1.89
    begin
    1.70
     beginning
    1.16
     zač
    1.11
     begins
    1.07
     begin
    1.05
     beginnings
    1.02
    beginning
    0.97
     began
    0.96
    Activations Density 0.010%

    No Known Activations