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 - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 966
    Prev
    Next
    INDEX
    Explanations

    references to communal dining experiences and interactions at the dining table

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

    elements related to family meals and international diplomacy

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

    dining table

    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-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
    urut
    -0.42
    mphony
    -0.41
    нци
    -0.39
    ově
    -0.38
    dor
    -0.37
    Orsay
    -0.36
    thor
    -0.36
    hor
    -0.36
    Handle
    -0.36
    handle
    -0.35
    POSITIVE LOGITS
     table
    2.23
     tables
    1.95
     Table
    1.88
    Table
    1.76
     Tables
    1.73
    table
    1.67
    TABLE
    1.66
     TABLE
    1.62
     tafel
    1.61
     TABLES
    1.55
    Activations Density 0.065%

    No Known Activations