© 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. 10982
    Prev
    Next
    INDEX
    Explanations

    numerical values and their relationships in scientific data

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

    text inside parentheses, particularly in scientific or technical documents where information is enclosed in parenthetical statements.

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

    formal scientific-report language that presents quantified results and measurements, with references to figures/tables/steps and occasional mathematical or Greek notation.

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

    parenthetical expressions or phrases that provide context, clarification, or technical specifications within scientific or technical documents.

    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
     I
    -0.35
    Odkazy
    -0.35
     II
    -0.34
     .
    -0.34
     L
    -0.33
     M
    -0.32
     H
    -0.32
     rozgry
    -0.32
     maar
    -0.32
    ToEnd
    -0.32
    POSITIVE LOGITS
     autorytatywna
    0.76
    Personendaten
    0.73
    +#+#
    0.71
    Билгалдахарш
    0.69
    makeConstraints
    0.68
     queſta
    0.66
     Roskov
    0.66
     للاسماء
    0.66
    ésultats
    0.66
     surla
    0.66
    Activations Density 0.576%

    No Known Activations