© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 7697
    Prev
    Next
    INDEX
    Explanations

    product and brand names, especially those related to chemicals, boats, and consumer goods.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @yassogba

    references to data structures and their components in programming contexts

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

    performance, Pod, opportunities, inmates, expansion, wife, Germany

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

    concrete content nouns denoting specific entities or tasks, especially within quoted instructions, lists, or technical examples.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     ویکی‌پدی
    -1.19
    EndGlobalSection
    -1.01
    featureID
    -1.00
    ьаж
    -0.97
     Administrativna
    -0.96
     noDo
    -0.91
    msgTypes
    -0.88
    IndentedString
    -0.85
    ंदीखरीदारी
    -0.82
    Personensuche
    -0.80
    POSITIVE LOGITS
     surla
    0.40
    extAlignment
    0.33
     Вікіпе
    0.32
     fortæ
    0.32
    SneakyThrows
    0.28
     rhestr
    0.28
     désolés
    0.27
     filós
    0.26
    ClientRect
    0.26
    󠁢
    0.26
    Activations Density 35.099%

    No Known Activations