© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 130198
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on the main “lead‐in” verbs and nouns that open a new section or highlight—e.g. the capitalized action words and subjects (This video demonstrates, Editing occurred, My highlight, The site includes, The Register itself) that introduce or headline a passage.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    linking words and descriptions

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    iod
    -0.92
     and
    -0.92
     "",
    -0.91
    ViewGroup
    -0.87
     when
    -0.87
     "','
    -0.86
    信心
    -0.85
     ഗ
    -0.84
    ございません
    -0.84
    ERVIEW
    -0.84
    POSITIVE LOGITS
     also
    2.22
     therefore
    1.63
     також
    1.48
     thus
    1.33
     также
    1.29
     inoltre
    1.24
     however
    1.22
     ayrıca
    1.19
     therefor
    1.10
     więc
    1.09
    Activations Density 0.159%

    No Known Activations