Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 56791
    Prev
    Next
    INDEX
    Explanations

    shebang lines and similar script header syntax

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

    shell script and Makefile elements

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
     '\\;'
    -0.83
     שוליים
    -0.81
     queſta
    -0.80
     للمعارف
    -0.80
     Weiſe
    -0.78
    <unused28>
    -0.77
    FunctionFlags
    -0.77
    <unused8>
    -0.77
    <unused3>
    -0.77
    [@BOS@]
    -0.77
    POSITIVE LOGITS
    #!/
    0.55
    CloseOperation
    0.39
     enver
    0.33
     
    0.31
     basan
    0.30
     /
    0.29
    Entra
    0.28
    ://
    0.28
     calibration
    0.26
    AUTHENT
    0.25
    Activations Density 0.001%

    No Known Activations