© 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. 70700
    Prev
    Next
    INDEX
    Explanations

    numbers and actions

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

    This neuron fires on content‐bearing (informative) tokens—i.e. rarer words, domain terms, and numbers—rather than on 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
    pfung
    -0.80
    فران
    -0.76
    ඬ
    -0.73
     adjournment
    -0.71
     macular
    -0.70
     aukš
    -0.70
     sessile
    -0.70
     utom
    -0.70
     Beng
    -0.69
    返信
    -0.68
    POSITIVE LOGITS
    nija
    0.89
     selected
    0.83
    atula
    0.78
     crescente
    0.77
    )}+\
    0.77
     sereia
    0.76
     lanterna
    0.75
     existente
    0.75
     exists
    0.75
    ्त
    0.74
    Activations Density 0.116%

    No Known Activations