© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 99099
    Prev
    Next
    INDEX
    Explanations

    number of deaths

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

    The neuron is strongly activated by commas and numerical/date tokens—that is, punctuation and numbers marking clause breaks or numeric information.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     و
    -1.85
    ফ
    -1.44
     vores
    -1.43
    čnosti
    -1.43
    jectable
    -1.43
     altezza
    -1.42
    thanol
    -1.41
    ternos
    -1.37
    hydrogenase
    -1.36
    zości
    -1.36
    POSITIVE LOGITS
     two
    1.74
    .
    1.56
    สอง
    1.45
     three
    1.29
     две
    1.27
    ヨーロ
    1.27
     several
    1.21
     during
    1.20
     jsem
    1.19
     used
    1.18
    Activations Density 0.122%

    No Known Activations