Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13985
    Prev
    Next
    INDEX
    Explanations

    HTML and XML-style tags and attributes, particularly those related to document type definitions (DTDs) and character encodings.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @myinzi123

    references to legal cases and citations

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

    technical terms, code elements, and references commonly found in programming, software licensing, and digital documentation.

    oai_token-act-pair · gpt-4.1-2025-04-14Triggered by @ywhp2c8tjm

    Software, Licenses, legal references

    np_acts-logits-general · gemini-2.0-flash

    software licenses and citations

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

    technical boilerplate and metadata from web/code documents—such as HTML/XML declarations, licensing/legal headers, URLs, and long alphanumeric/hash strings

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
    <bos>
    -0.66
    The
    -0.39
    <
    -0.36
    A
    -0.35
    <strong>
    -0.34
    [
    -0.33
    I
    -0.32
    1
    -0.32
    (
    -0.32
     and
    -0.31
    POSITIVE LOGITS
    ConstraintMaker
    0.74
     ब्रेकडाउन
    0.72
     autorytatywna
    0.69
    httphttps
    0.69
    KommentareTeilen
    0.68
    ScopeManager
    0.68
     estekak
    0.67
    tagHelperRunner
    0.67
     CreateTagHelper
    0.67
    MLLoader
    0.67
    Activations Density 30.462%

    No Known Activations