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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 36-GEMMASCOPE-RES-16K
    6. 14731
    Prev
    Next
    INDEX
    Explanations

    instances of the word "ever" and its variations, indicating a focus on the concept of time or permanence

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.36.hook_resid_post
    Hook Layer
    36
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     altogether
    -0.41
    Redacción
    -0.37
     Hague
    -0.36
     giustizia
    -0.36
    一直
    -0.36
     Ignorance
    -0.35
    phas
    -0.35
    着丈
    -0.34
     geography
    -0.34
     giả
    -0.34
    POSITIVE LOGITS
    greens
    0.74
    last
    0.70
    glades
    0.63
    expandindo
    0.62
    lasting
    0.61
    present
    0.60
    UrlResolution
    0.59
    thing
    0.58
    ready
    0.55
     wondered
    0.54
    Activations Density 0.056%

    No Known Activations