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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 68926
    Prev
    Next
    INDEX
    Explanations

    phrases indicating how something is constructed or created

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

    what something is made of

    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
    近平
    -0.57
    beqa
    -0.57
     frequência
    -0.55
    ompi
    -0.54
    agues
    -0.53
     EventArgs
    -0.52
     חיצוניים
    -0.52
    ague
    -0.52
    $_['
    -0.52
     henkil
    -0.52
    POSITIVE LOGITS
    Made
    0.62
    MADE
    0.56
     manufacture
    0.56
    made
    0.55
     fabriqué
    0.55
     Made
    0.52
     manufactured
    0.52
     fabriquer
    0.50
     made
    0.49
    Manufacture
    0.49
    Activations Density 0.011%

    No Known Activations