© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 127750
    Prev
    Next
    INDEX
    Explanations

    names and specific terms

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

    This neuron responds to capitalized tokens, i.e. words that are part of proper names or named entities.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    1
    -2.91
    \
    -2.73
    "
    -2.56
    tr
    -2.23
     Their
    -2.19
    3
    -2.16
    lo
    -2.13
    Q
    -2.13
     MUCH
    -2.11
     Öffnungszeiten
    -2.11
    POSITIVE LOGITS
    
    2.73
    ￬
    2.56
    ቍ
    2.55
    
    2.25
     he
    2.22
    providedIn
    2.22
    埽
    2.19
     них
    2.17
    sedown
    2.16
    __);
    2.14
    Activations Density 0.002%

    No Known Activations