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. 20-GEMMASCOPE-RES-16K
    6. 7727
    Prev
    Next
    INDEX
    Explanations

    arguments and assertions made in legal contexts

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

    arguments or assertions

    np_acts-logits-general · gemini-2.0-flash

    arguing or asserting points

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    DeleteBehavior
    -0.65
    urlpatterns
    -0.64
    Rüyada
    -0.63
    CodeAttribute
    -0.60
     Meksiku
    -0.60
    Curios
    -0.57
    VersionUID
    -0.57
     estekak
    -0.57
    dotenv
    -0.57
    出版年
    -0.56
    POSITIVE LOGITS
     arguments
    2.33
     argument
    2.31
     Arguments
    2.03
     argue
    1.95
     Argument
    1.93
    arguments
    1.92
     argued
    1.92
    argument
    1.90
     argumento
    1.89
     arguing
    1.87
    Activations Density 1.220%

    No Known Activations