© 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. 10-GEMMASCOPE-RES-131K
    7. 41479
    Prev
    Next
    INDEX
    Explanations

    defeated, beaten, outdone

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

    The neuron fires on numeric tokens and measurements (integer or decimal values) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    statements describing a side being defeated or outmatched—often framed as “X was beaten/outscored by Y” and frequently accompanied by scores or margins.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    the phrase "beaten by" or "defeated by" in sports and competitive contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered 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
     mesmas
    -1.46
    already
    -1.37
    :
    -1.36
     Organizador
    -1.30
    demás
    -1.30
     difficiles
    -1.30
     Trabaj
    -1.29
     gradualmente
    -1.28
     jednocześnie
    -1.26
     nouvel
    -1.23
    POSITIVE LOGITS
     one
    1.54
    ttino
    1.36
    ()]
    1.33
    ÑO
    1.32
    ʐ
    1.30
     ONE
    1.29
    dibility
    1.29
    ệu
    1.25
    calyp
    1.20
    OBSERV
    1.18
    Activations Density 0.030%

    No Known Activations