© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 61424
    Prev
    Next
    INDEX
    Explanations

    script or code blocks

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

    The neuron fires on code markup tokens—especially template include directives and HTML/JavaScript tags (e.g. `{include file=…}`, `<script>`, angle brackets, etc.).

    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
     çünkü
    -1.41
     rigue
    -1.39
     silnika
    -1.37
     españa
    -1.35
    aktionen
    -1.34
    izaron
    -1.34
     להם
    -1.31
    geführt
    -1.30
     boucles
    -1.30
     ?
    -1.29
    POSITIVE LOGITS
     those
    1.71
     but
    1.58
    ścioła
    1.44
    After
    1.38
    Although
    1.36
     Wartawan
    1.34
     with
    1.31
     Although
    1.29
    Those
    1.28
    exprimer
    1.27
    Activations Density 0.006%

    No Known Activations