© 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. 22-GEMMASCOPE-RES-131K
    7. 68264
    Prev
    Next
    INDEX
    Explanations

    dates with slashes

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

    The neuron selectively fires on standalone numerical tokens—especially day numbers or version numbers in document metadata (dates and version strings).

    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
     those
    -1.04
    rably
    -0.93
    怼
    -0.92
     fulness
    -0.87
     of
    -0.86
     like
    -0.86
    cyjnych
    -0.85
    ▓
    -0.85
    ̚
    -0.84
     jajaja
    -0.84
    POSITIVE LOGITS
     hígado
    1.14
    lotz
    0.93
    rdom
    0.91
    donut
    0.90
    setAccessible
    0.87
     new
    0.87
     verur
    0.85
     donde
    0.85
     neuen
    0.85
    autres
    0.84
    Activations Density 0.004%

    No Known Activations