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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 5163
    Prev
    Next
    INDEX
    Explanations

    terms related to studies or analyses involving measurements and their implications

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    numbers and numerical measurements, especially in scientific or medical contexts.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    references to documents, studies, or written materials being discussed in academic or analytical contexts.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    key content nouns—especially abstract or technical head nouns denoting concepts, conditions, measurements, or agreements.

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

    specialized terms or phrases from technical, scientific, medical, or formal domains.

    oai_token-act-pair · deepseek-r1Triggered by @jyhe0408

    words that appear near the end of academic or technical documents, particularly in scientific studies and formal analyses.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    uxxxx
    -0.80
     propOrder
    -0.77
     pinulongan
    -0.73
     intStringLen
    -0.71
     Infórmanos
    -0.71
     ErrIntOverflow
    -0.68
    PreferredItem
    -0.68
     ویکی‌پدی
    -0.68
     bezeichneter
    -0.67
    istoitu
    -0.67
    POSITIVE LOGITS
     will
    0.35
    fair
    0.35
     
    0.35
    ,
    0.34
    .
    0.34
     Ordnung
    0.33
     keş
    0.32
    zusetzen
    0.31
    fra
    0.31
     '
    0.31
    Activations Density 0.276%

    No Known Activations