© 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. 22-GEMMASCOPE-RES-131K
    7. 67645
    Prev
    Next
    INDEX
    Explanations

    state names and country names

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

    This neuron fires on the semantically “contentful” words—rare or domain-specific nouns, verbs, and adjectives (e.g. titles like JUDGE or CLERK, instruction verbs like Let or Obtain, technical terms, etc.)—while ignoring common function words.

    oai_token-act-pair · o4-miniTriggered 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
    Not in Any Lists

    No Comments

    Negative Logits
    érales
    -0.89
    栊
    -0.88
    <strong>
    -0.84
    ARGS
    -0.82
    ఝ
    -0.82
    FIX
    -0.80
     الجديد
    -0.79
     ซ
    -0.77
    *******
    -0.77
     things
    -0.77
    POSITIVE LOGITS
     inappropriés
    0.96
    ménage
    0.96
     will
    0.90
     gangen
    0.88
    denim
    0.88
    ͨ
    0.88
    彼らは
    0.86
    idän
    0.86
    Dossier
    0.86
     gyere
    0.85
    Activations Density 0.006%

    No Known Activations