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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 5305
    Prev
    Next
    INDEX
    Explanations

    attends to contextual indicators of outcomes or future events from earlier tokens that mention specific teams or positions

    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
    0:0.10
    1:0.13
    2:0.11
    3:0.12
    4:0.13
    5:0.05
    6:0.15
    7:0.17
    Negative Logits
     ་་
    -0.30
     Shakspeare
    -0.30
    addContainerGap
    -0.30
     itſelf
    -0.30
    isseurs
    -0.29
     Ingo
    -0.29
     myſelf
    -0.29
    RegistryLite
    -0.29
     PIB
    -0.26
     Houſe
    -0.26
    POSITIVE LOGITS
     ligiloj
    0.37
    İstinadlar
    0.32
    Havolalar
    0.31
    PMailer
    0.30
    Билгалдахарш
    0.29
    RefreshLayout
    0.29
     suivants
    0.28
    Autoritní
    0.28
     earlier
    0.27
     lenker
    0.27
    Activations Density 0.047%

    No Known Activations