© 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. 39624
    Prev
    Next
    INDEX
    Explanations

    common words and trailing letters

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

    The neuron strongly activates on contentful, domain‐specific tokens (i.e. meaningful nouns, verbs, and adjectives) and stays off for 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
     ContentValues
    -1.04
    Географија
    -1.02
    Gön
    -0.97
    (≧
    -0.93
    nikov
    -0.93
     okru
    -0.91
     getArguments
    -0.91
     ambas
    -0.89
     akrab
    -0.88
    archiw
    -0.87
    POSITIVE LOGITS
     einsetzen
    0.89
    a
    0.87
    d
    0.83
    o
    0.82
    的白
    0.81
     contiennent
    0.81
    нести
    0.79
     تنها
    0.79
    igator
    0.79
     italics
    0.78
    Activations Density 0.077%

    No Known Activations