© 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. 128102
    Prev
    Next
    INDEX
    Explanations

    materials and methods

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

    technical terms related to computer science, tests, and academic disciplines.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron fires on multi‐letter, content‐bearing tokens—i.e. domain-specific keywords or section headers—rather than on function words or short/common tokens.

    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
     rése
    -1.88
     dekol
    -1.88
     mikrofon
    -1.80
     kompres
    -1.77
     fidé
    -1.77
     themſelves
    -1.73
    those
    -1.73
     kokos
    -1.70
    ⍢
    -1.69
     reklam
    -1.68
    POSITIVE LOGITS
     only
    2.52
     when
    2.20
     will
    2.19
     was
    2.16
    </b>
    2.09
     is
    2.08
     at
    2.08
     an
    2.06
     in
    2.02
     a
    1.93
    Activations Density 0.143%

    No Known Activations