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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 0-GEMMASCOPE-RES-16K
    6. 3330
    Prev
    Next
    INDEX
    Explanations

    the word "separation" and similar words, likely indicating a focus on divisions or splits of some kind.

    oai_token-act-pair · gemini-1.5-flashTriggered by @johnny

    The marker "separation" is consistently used to encapsulate key concepts or points of focus in the text, especially related to distinct categories, differences, or the act of dividing or separating something. This consistent pattern suggests a recurring emphasis on identifying separations or distinctions within various contexts across the examples.

    eleuther_acts_top20 · gpt-3.5-turboTriggered by @hijohnnylinmonorepo

    terms related to separation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_0/width_16k/average_l0_105
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.hook_resid_post
    Hook Layer
    0
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    errHandler
    -0.73
    Portail
    -0.72
     AssemblyCulture
    -0.70
    complexContent
    -0.69
    DeleteBehavior
    -0.67
    TagMode
    -0.65
    WebServlet
    -0.64
    RuleContext
    -0.62
    ModelSerializer
    -0.61
     GetComponent
    -0.58
    POSITIVE LOGITS
     terrace
    1.03
     spacing
    0.84
    Spacing
    0.80
     terraces
    0.79
     merits
    0.78
     Spacing
    0.76
     balcony
    0.75
     terrasse
    0.74
    spacing
    0.73
     Balcony
    0.72
    Activations Density 0.074%

    No Known Activations