© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 38366
    Prev
    Next
    INDEX
    Explanations

    source/date/classification

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

    The neuron activates on numeric digit sequences, i.e. numbers embedded in the text.

    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
    וּ
    -1.17
    ִי
    -1.10
    וֹ
    -1.10
    ële
    -1.08
    kenen
    -1.08
    њу
    -1.07
    ޤ
    -1.06
    znego
    -1.05
    ungkapkan
    -1.05
    _->
    -1.05
    POSITIVE LOGITS
    It
    1.15
    ientôt
    1.10
     watching
    1.04
     wasn
    1.04
    ratulations
    1.03
     område
    1.02
    They
    1.01
     этой
    1.01
     almacenar
    1.00
     correctly
    1.00
    Activations Density 0.001%

    No Known Activations