© 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. 25847
    Prev
    Next
    INDEX
    Explanations

    titles and beginnings of text

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

    The neuron detects XML-style documentation comment markers (the “///” tokens) in source code.

    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
     all
    -1.52
     what
    -1.51
     even
    -1.45
     every
    -1.41
     its
    -1.41
     each
    -1.36
     before
    -1.35
     without
    -1.34
     any
    -1.34
     where
    -1.33
    POSITIVE LOGITS
     grotte
    1.32
     manier
    1.30
     maniere
    1.29
     ähnliche
    1.26
     montagne
    1.16
     vitamine
    1.16
     schuhe
    1.16
    poons
    1.15
    </em>
    1.14
     truff
    1.13
    Activations Density 0.000%

    No Known Activations