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. 1163
    Prev
    Next
    INDEX
    Explanations

    proper nouns, especially names

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

    instances of the name "Mazurek"

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

    Mazurkiewicz, Hirata, CR1

    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
    weit
    -0.57
    óis
    -0.56
    ̀nh
    -0.55
    voort
    -0.54
     segreto
    -0.53
    checkable
    -0.52
     Lacy
    -0.52
    mstyle
    -0.51
    atuur
    -0.50
    iors
    -0.50
    POSITIVE LOGITS
    Maz
    1.07
     Maz
    1.00
     MAZ
    0.94
    SBATCH
    0.83
     maz
    0.77
    maz
    0.75
    UserScript
    0.73
    Ligações
    0.68
     AssemblyTitle
    0.65
     NUKAT
    0.64
    Activations Density 0.015%

    No Known Activations