© 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. 66578
    Prev
    Next
    INDEX
    Explanations

    names, dates, or historical associations

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

    The neuron strongly activates on parts of capitalized personal names, effectively detecting proper names of people.

    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
     will
    -0.91
     causes
    -0.90
     this
    -0.88
     epitom
    -0.84
     last
    -0.83
    第四
    -0.82
    фика
    -0.82
    Last
    -0.82
    ضية
    -0.82
     new
    -0.82
    POSITIVE LOGITS
    <0x80>
    1.15
     heutigen
    1.07
     heutige
    1.01
     нынеш
    1.01
    eventually
    0.98
    后来
    0.98
     eventual
    0.98
    
    0.94
     grew
    0.94
     générations
    0.94
    Activations Density 0.097%

    No Known Activations