© 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. 22-GEMMASCOPE-RES-131K
    7. 15200
    Prev
    Next
    INDEX
    Explanations

    meeting notes or correspondence

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

    The neuron strongly activates on numeric tokens—especially multi‐digit sequences and dates—highlighting numbers (e.g. day, month, and year components) 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
     vanil
    -1.41
     koste
    -1.35
     potret
    -1.34
     maintenant
    -1.32
     статьи
    -1.31
     gede
    -1.31
     kompeti
    -1.30
    éroport
    -1.29
    According
    -1.27
    "><!--
    -1.26
    POSITIVE LOGITS
     regarding
    1.92
     later
    1.88
     shortly
    1.73
     after
    1.72
     during
    1.66
     several
    1.56
     about
    1.54
     on
    1.52
     immediately
    1.51
     from
    1.45
    Activations Density 0.006%

    No Known Activations