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. 6034
    Prev
    Next
    INDEX
    Explanations

    attends to the token "async" from function-related tokens

    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
    0:0.23
    1:0.19
    2:0.15
    3:0.11
    4:0.04
    5:0.04
    6:0.07
    7:0.14
    Negative Logits
    slidesPer
    -0.29
    Aiheesta
    -0.27
    imentel
    -0.27
     McQu
    -0.26
    
    -0.25
    ToAction
    -0.25
    componentWill
    -0.24
    chtend
    -0.24
    地看着
    -0.23
    SpringBoot
    -0.23
    POSITIVE LOGITS
     فريبيس
    0.45
     Савезне
    0.38
     purpoſe
    0.37
     Italijanski
    0.35
    styleType
    0.34
     Infórmanos
    0.33
     pleaſure
    0.33
     myſelf
    0.33
     biling
    0.31
     fevere
    0.31
    Activations Density 0.020%

    No Known Activations