© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 16583
    Prev
    Next
    INDEX
    Explanations

    Copyright, Late, end, agreed

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

    The neuron is sensitive to rare or standout tokens—uncommon words, punctuation marks, numbers, or jargon—that tend to carry high specificity in a document.

    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_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    there
    -1.54
     it
    -1.51
     می‌گو
    -1.48
    UCIÓN
    -1.36
    "—
    -1.36
    CCIÓN
    -1.27
    !—
    -1.22
     principally
    -1.20
     because
    -1.18
     mostly
    -1.18
    POSITIVE LOGITS
     logiciels
    1.39
     graphiques
    1.38
     ï
    1.37
     Paramètres
    1.33
     دقی
    1.30
     Réalisation
    1.27
     acestui
    1.26
    <strong>
    1.25
     Outils
    1.25
     Ajouter
    1.24
    Activations Density 0.005%

    No Known Activations