© 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. 20264
    Prev
    Next
    INDEX
    Explanations

    punctuation followed by conjunctions

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

    The neuron fires most strongly on the formal, “official” signposts and modal or evidential verbs you find in legal- or academic-style writing (e.g. the commas delimiting header metadata and words like “should,” “represents,” “demonstrated,” “noticed,” 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
     this
    -1.91
     it
    -1.78
     our
    -1.55
    の通り
    -1.47
    好事
    -1.41
    ień
    -1.41
    Ꮗ
    -1.40
     the
    -1.32
     Vorteil
    -1.32
    тна
    -1.30
    POSITIVE LOGITS
    ”)
    2.22
    ”;
    1.91
    ”).
    1.75
    ”
    1.74
    四个
    1.69
    ...”
    1.69
    …”
    1.66
    龵
    1.63
    埽
    1.63
    ”،
    1.63
    Activations Density 0.459%

    No Known Activations