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 - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 1603
    Prev
    Next
    INDEX
    Explanations

    descriptions of flavors and qualities in food products, particularly oils

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

    Describing a flavor or taste

    np_acts-logits-general · gemini-2.0-flash

    flavor, taste, smell

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    HtmlAttribute
    -0.72
    出版年
    -0.66
    AxisAlignment
    -0.65
    DockStyle
    -0.62
    IntoConstraints
    -0.62
    XmlAccessorType
    -0.60
    題
    -0.57
    GeneratedCode
    -0.57
    mtrl
    -0.56
    ArgumentParser
    -0.56
    POSITIVE LOGITS
     flavor
    1.17
     taste
    1.08
     flavour
    1.05
     flavors
    1.03
     tastes
    0.97
     flavours
    0.94
     Flavor
    0.94
    flavor
    0.89
    Flavor
    0.87
     subtle
    0.81
    Activations Density 0.099%

    No Known Activations