© 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. 34-GEMMASCOPE-RES-131K
    7. 49356
    Prev
    Next
    INDEX
    Explanations

    numbers, web urls, and word beginnings like 'respon'

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

    This neuron detects numeric tokens (individual digits and numbers, especially in dates and counts).

    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
    大致
    -1.16
     wohin
    -1.13
     a
    -1.12
     acestea
    -1.10
    mainly
    -1.05
     solches
    -1.01
    these
    -0.97
     these
    -0.97
    Leírás
    -0.94
     geralmente
    -0.93
    POSITIVE LOGITS
     the
    1.86
    ˃
    1.20
    chè
    1.15
    aneamente
    1.14
     Février
    1.13
     pembentukan
    1.13
     italiano
    1.11
     pacientes
    1.01
    ???:
    0.99
    椋
    0.99
    Activations Density 0.081%

    No Known Activations