© 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. 115738
    Prev
    Next
    INDEX
    Explanations

    prob clues

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

    This neuron fires on tokens that look like programming or markup elements—things like punctuation and symbols in code (e.g. hyphens in identifiers, apostrophes, file‐type words like “image,” “inc,” “radius,” etc.) rather than ordinary natural‐language words.

    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
    ridae
    -0.82
    LTC
    -0.71
    ronomy
    -0.70
     előtt
    -0.70
    erry
    -0.69
     nær
    -0.69
     Samaria
    -0.69
    jercicio
    -0.69
     Pente
    -0.68
     poursuivre
    -0.68
    POSITIVE LOGITS
     tane
    0.75
     probar
    0.73
     espan
    0.69
    Ownership
    0.69
     cero
    0.68
     оли
    0.66
     prob
    0.66
     Ownership
    0.66
     Diciembre
    0.66
     provas
    0.66
    Activations Density 0.034%

    No Known Activations