© 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. 84106
    Prev
    Next
    INDEX
    Explanations

    Q: or See

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

    The neuron fires on numeric tokens—especially sequences of digits and decimal numbers—i.e. it detects numeric literals 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
     aldri
    -1.37
     Kỳ
    -1.30
     zelve
    -1.29
    וּ
    -1.24
    rodz
    -1.24
     påvir
    -1.17
     spesielt
    -1.11
     vært
    -1.11
     zask
    -1.11
     abstrait
    -1.10
    POSITIVE LOGITS
    另一方面
    1.07
    ph
    1.03
    : 
    1.02
    p
    1.01
    ep
    0.98
    em
    0.97
    cl
    0.96
    L
    0.96
     our
    0.94
     The
    0.94
    Activations Density 0.001%

    No Known Activations