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. 45947
    Prev
    Next
    INDEX
    Explanations

    sequences or patterns of asterisks and underscores, possibly indicating emphasis or special notation in the text

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

    high occupation, commitment, screened, non, dynamic, Feynman, sources, high, rms

    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
    SequentialGroup
    -0.54
    principalTable
    -0.54
    FormTagHelper
    -0.54
    LayoutStyle
    -0.51
    UpInside
    -0.49
     ferons
    -0.48
     المعيارى
    -0.48
    CPtr
    -0.47
    ständen
    -0.47
    MethodManager
    -0.46
    POSITIVE LOGITS
    (!)
    0.54
     (!)
    0.50
     चीज़ों
    0.43
    <i>
    0.42
    !
    0.42
    клопе
    0.41
     jsPsych
    0.40
    !)
    0.39
    <em>
    0.39
    oderm
    0.37
    Activations Density 0.028%

    No Known Activations