© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 3680
    Prev
    Next
    INDEX
    Explanations

    programming syntax and control structures

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

    programming language syntax and code structures, especially in JavaScript or Java.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    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
    ########.
    -0.69
    RenderAtEndOf
    -0.66
     defaultstate
    -0.61
    таратура
    -0.60
     linkovi
    -0.60
     BytesLike
    -0.59
    posedge
    -0.57
     kasarigan
    -0.57
    IntOverflow
    -0.55
     queſta
    -0.55
    POSITIVE LOGITS
    AddHtmlAttribute
    0.52
     no
    0.41
     although
    0.38
     vollen
    0.38
     Spitze
    0.37
     immediately
    0.36
     therefore
    0.35
     holding
    0.35
    감
    0.34
     gradually
    0.34
    Activations Density 0.067%

    No Known Activations