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-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 754
    Prev
    Next
    INDEX
    Explanations

    structured data related to management or core systems

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

    numerical values and references related to management or organizational terminology

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    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
    Citation
    -0.58
     AppModule
    -0.54
    jstor
    -0.49
    Autoritní
    -0.48
    ktör
    -0.47
     isSet
    -0.47
    pixabay
    -0.46
    Activité
    -0.46
     turnovers
    -0.46
    antwort
    -0.45
    POSITIVE LOGITS
     versions
    0.74
     version
    0.73
    flavor
    0.72
     flavor
    0.71
     flavour
    0.71
     purpoſe
    0.71
     Jefus
    0.70
    latest
    0.69
     transfieras
    0.68
     itſelf
    0.68
    Activations Density 0.027%

    No Known Activations