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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Attention Out - 16k
    5. 10-GEMMASCOPE-ATT-16K
    6. 6234
    Prev
    Next
    INDEX
    Explanations

    attends to related tokens specifying types or categories, such as "leaf," "cell," "child," and "lipid," from tokens that signify the same type but are presented in a different context, such as "leaf," "cell," "child," or "fat."

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k
    Configuration
    google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.10.attn.hook_z
    Hook Layer
    10
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.05
    1:0.17
    2:0.05
    3:0.04
    4:0.12
    5:0.44
    6:0.04
    7:0.05
    Negative Logits
    DockStyle
    -0.41
    AutoScale
    -0.39
     AssemblyCulture
    -0.38
    $_['
    -0.35
    orteur
    -0.34
    CloseOperation
    -0.34
    mybatisplus
    -0.34
     Folge
    -0.33
    ChildScrollView
    -0.32
    ябре
    -0.32
    POSITIVE LOGITS
     referenties
    0.30
    ']))
    
    0.27
     Cik
    0.27
     Adair
    0.27
    redientes
    0.27
    ly
    0.27
    }))
    
    0.27
     generali
    0.26
    Lage
    0.26
     Sankt
    0.26
    Activations Density 0.409%

    No Known Activations