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-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 40708
    Prev
    Next
    INDEX
    Explanations

    special character patterns and numerical data related to structures or entities in a dataset

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

    Values within table cells, particularly numerical data or category labels in data tables formatted with columns and row separators, often appearing in scientific/academic publications.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered 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
    Not in Any Lists

    No Comments

    Negative Logits
    RectangleBorder
    -0.81
     AssemblyCulture
    -0.72
    :✨
    -0.70
    PerformLayout
    -0.63
     *++
    -0.63
    Personensuche
    -0.61
    antası
    -0.60
     MainAxisSize
    -0.59
    DockStyle
    -0.59
    onViewCreated
    -0.57
    POSITIVE LOGITS
    __(/*!
    0.40
    EM
    0.37
     pin
    0.35
     only
    0.35
     
    0.34
     concerned
    0.34
     nisso
    0.33
     room
    0.33
     floor
    0.33
    대로
    0.32
    Activations Density 0.089%

    No Known Activations