© 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. 58905
    Prev
    Next
    INDEX
    Explanations

    sports and competition

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

    The neuron strongly activates on tokens that serve as titles, roles, or proper‐name descriptors—i.e. words naming people’s occupations, honors, works, or categories.

    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
     garantis
    -0.87
     Meksiko
    -0.85
     rumore
    -0.84
     unga
    -0.76
     inspiri
    -0.75
     busto
    -0.72
     twa
    -0.72
    ちょっと
    -0.71
     alha
    -0.70
     alban
    -0.67
    POSITIVE LOGITS
     football
    2.94
     basketball
    2.92
     soccer
    2.88
     tennis
    2.61
     golf
    2.58
     sports
    2.58
     baseball
    2.42
     hockey
    2.42
     rugby
    2.41
     volleyball
    2.41
    Activations Density 0.188%

    No Known Activations