© 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. 34-GEMMASCOPE-RES-131K
    7. 12279
    Prev
    Next
    INDEX
    Explanations

    memory management and properties

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

    This neuron fires most strongly on low‐frequency or distinctive tokens (e.g. uncommon words and identifiers) rather than on common 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_34/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
    ようになりました
    -0.84
     skola
    -0.81
     Minor
    -0.81
     chansons
    -0.79
     début
    -0.77
     refuser
    -0.77
     cepat
    -0.77
    мышлен
    -0.75
    鲍
    -0.75
     disruptions
    -0.73
    POSITIVE LOGITS
     caches
    0.88
    requestCode
    0.83
    ursales
    0.82
    ThemeProvider
    0.82
     vaše
    0.81
     memory
    0.79
    Kommentare
    0.79
    InMemory
    0.78
    alloc
    0.78
     cache
    0.78
    Activations Density 0.001%

    No Known Activations