© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 75256
    Prev
    Next
    INDEX
    Explanations

    punctuation and ending 's

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

    The neuron strongly activates on very frequent function words and common English tokens (e.g. articles, pronouns, basic verbs, conjunctions).

    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
     vengono
    -1.12
    Alguns
    -1.00
    Some
    -0.96
    There
    -0.87
    应
    -0.85
     dazzling
    -0.83
     verranno
    -0.83
    uchte
    -0.82
    When
    -0.82
     verrà
    -0.82
    POSITIVE LOGITS
     its
    1.22
     ulo
    1.02
     prace
    0.98
    complete
    0.98
     และ
    0.97
    其
    0.96
     komplet
    0.95
     talks
    0.93
    みたい
    0.92
    specifically
    0.91
    Activations Density 0.066%

    No Known Activations