© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 6-GEMMASCOPE-RES-16K
    6. 14077
    Prev
    Next
    INDEX
    Explanations

    the presence of specific coding markers indicating the beginning of a new data block or section

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

    beginning of document tokens, particularly detecting the start of new documents.

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

    the beginning of a document or text (<bos> token), with particularly high activation values.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 6-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_6/width_16k/average_l0_70
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.6.hook_resid_post
    Hook Layer
    6
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     AspNetCore
    -1.05
     للمعارف
    -0.95
     snippetHide
    -0.92
     surla
    -0.82
     ComVisible
    -0.81
     protoimpl
    -0.80
    ValueStyle
    -0.77
     كومونز
    -0.75
    XmlAccessType
    -0.73
    Демографія
    -0.72
    POSITIVE LOGITS
    ↵↵
    0.90
    ↵
    0.85
    <eos>
    0.73
    <h2>
    0.67
    <tr>
    0.66
    ↵↵↵
    0.65
    ↵↵↵↵
    0.61
    section
    0.61
    </table>
    0.56
    The
    0.54
    Activations Density 0.022%

    No Known Activations