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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 6487
    Prev
    Next
    INDEX
    Explanations

    attends to numerical tokens that are followed by tokens that contain square brackets

    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
    0:0.07
    1:0.09
    2:0.07
    3:0.10
    4:0.08
    5:0.08
    6:0.32
    7:0.15
    Negative Logits
     saites
    -0.52
    DebuggerNonUser
    -0.42
     estekak
    -0.41
     beginnetje
    -0.41
    aarrggbb
    -0.40
    matchCondition
    -0.40
    tonsoft
    -0.40
    Parcelize
    -0.39
    Personensuche
    -0.39
    '&:
    -0.39
    POSITIVE LOGITS
    forName
    0.28
    hips
    0.27
     scissors
    0.26
    hog
    0.26
     Filed
    0.26
     ftu
    0.26
    illères
    0.24
    FT
    0.24
     Kiev
    0.24
     osu
    0.24
    Activations Density 0.001%

    No Known Activations