© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 3417
    Prev
    Next
    INDEX
    Explanations

    brand new

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

    The neuron is primarily responding to mentions of “new” or the lack thereof (“new,” “nothing new,” “what’s new,” “she’s not new,” “brand new”), i.e. references to novelty.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    unstable
    -0.77
    ogene
    -0.75
     brighter
    -0.74
    スン
    -0.72
     giusta
    -0.71
     Gesund
    -0.70
     harder
    -0.70
     ahora
    -0.69
    taller
    -0.69
     another
    -0.69
    POSITIVE LOGITS
     new
    2.19
     brand
    1.45
    new
    1.45
    新
    1.35
    ใหม่
    1.34
    NEW
    1.32
     NEW
    1.25
    新品
    1.25
    brand
    1.15
     baru
    1.15
    Activations Density 0.063%

    No Known Activations