© 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. 68205
    Prev
    Next
    INDEX
    Explanations

    description of issues

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

    The neuron is picking out unusually long word‐tokens (i.e. tokens whose character length exceeds a certain threshold).

    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
     It
    -2.55
    .
    -2.34
     
    -2.34
    \
    -2.20
    <bos>
    -2.20
    -
    -2.02
     ".
    -1.98
    thschild
    -1.95
     '
    -1.91
    馀
    -1.86
    POSITIVE LOGITS
     the
    3.19
     of
    2.73
     them
    2.50
     libur
    2.30
    栻
    2.28
     kenal
    2.20
     berge
    2.17
     gustado
    2.16
    Истори
    2.13
     menyen
    2.13
    Activations Density 0.057%

    No Known Activations