© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 9075
    Prev
    Next
    INDEX
    Explanations

    Schultz, Otis, Reginald, Marino, armature, laryngeal, testicular

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

    last names and proper nouns, with higher activation for names like "Schultz," "Otis," "Labrador," and other similar surnames.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron activates strongly on unusually long single tokens (especially multi‐syllable words or proper names).

    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
    .
    -3.52
    s
    -3.14
    (
    -3.08
    {
    -2.73
    訫
    -2.55
    м
    -2.52
    THE
    -2.39
    :
    -2.34
     namun
    -2.33
    <b>
    -2.23
    POSITIVE LOGITS
    觔
    2.66
     divertidas
    2.58
    蟜
    2.56
    醅
    2.48
     sufren
    2.47
    鉏
    2.45
    戬
    2.42
    鹇
    2.42
    骝
    2.42
     complicada
    2.39
    Activations Density 0.075%

    No Known Activations