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. 0-GEMMASCOPE-ATT-16K
    6. 6062
    Prev
    Next
    INDEX
    Explanations

    attends to tokens related to "Super" from tokens related to specific named concepts or categories

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k
    Configuration
    google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.attn.hook_z
    Hook Layer
    0
    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
    0:0.42
    1:0.01
    2:0.02
    3:0.01
    4:0.32
    5:0.13
    6:0.01
    7:0.04
    Negative Logits
     "..\..\
    -0.46
    :✨
    -0.39
     InputDecoration
    -0.34
    zemu
    -0.34
     zů
    -0.33
     Majority
    -0.32
     Réponses
    -0.32
     lät
    -0.32
     "..\..\..\
    -0.32
    šinou
    -0.31
    POSITIVE LOGITS
    
    0.33
    byshev
    0.30
    prepareStatement
    0.30
    ChildScrollView
    0.29
    Hentet
    0.28
    Suara
    0.28
    tvguidetime
    0.28
    ADELPHIA
    0.27
    "?>
    0.26
    createCanvas
    0.26
    Activations Density 0.022%

    No Known Activations