© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 104897
    Prev
    Next
    INDEX
    Explanations

    the beginning of various document sections or paragraphs

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot

    A special token "<<<bos>>>" that serves as a boundary marker between different text segments or documents in a corpus.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219
    Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    Negative Logits
     EconPapers
    -0.69
    <unused52>
    -0.57
    <unused28>
    -0.56
    <unused68>
    -0.56
    <unused8>
    -0.56
    [@BOS@]
    -0.56
    <unused79>
    -0.56
    <unused14>
    -0.55
    <unused41>
    -0.55
    <unused3>
    -0.55
    POSITIVE LOGITS
    DeleteBehavior
    0.41
     propOrder
    0.40
     Chwiliwch
    0.36
    SuspendLayout
    0.34
     transférez
    0.33
     Савезне
    0.32
    astéroïdes
    0.31
    تقاوى
    0.31
    पया
    0.31
    LookAnd
    0.29
    Activations Density 0.063%

    No Known Activations