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. 10
    Prev
    Next
    INDEX
    Explanations

    attends to tokens that imply movement or progression from corresponding tokens marking a specific state or position

    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
    0:0.72
    1:0.01
    2:0.02
    3:0.01
    4:0.09
    5:0.05
    6:0.02
    7:0.03
    Negative Logits
     initComponents
    -0.32
    NOPQRST
    -0.30
    UrlResolution
    -0.27
    ToBounds
    -0.26
    SequentialGroup
    -0.26
    queryInterface
    -0.26
     resourceCulture
    -0.26
     ujednoznacz
    -0.25
    Oneof
    -0.25
     marmor
    -0.24
    POSITIVE LOGITS
    zkiem
    0.32
    ukan
    0.32
    iteurs
    0.29
    styleType
    0.29
    lgari
    0.28
    felves
    0.28
    ască
    0.27
    Slf
    0.27
    tedly
    0.27
    DoubleQuotes
    0.26
    Activations Density 0.005%

    No Known Activations