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-2-9B
    4. Attention Out - 16k
    5. 0-GEMMASCOPE-ATT-16K
    6. 1271
    Prev
    Next
    INDEX
    Explanations

    attends to the leading comment token from specific functionality tokens

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61
    Prompts (Dashboard)
    16,384 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
    0:0.02
    1:0.01
    2:0.01
    3:0.01
    4:0.01
    5:0.02
    6:0.20
    7:0.38
    8:0.04
    9:0.03
    10:0.01
    11:0.01
    12:0.02
    13:0.03
    14:0.10
    15:0.03
    Negative Logits
    SharedDtor
    -0.31
    /**
    -0.29
    OGND
    -0.29
    
    -0.27
     vägen
    -0.26
     <=",
    -0.26
     مشين
    -0.26
     ExecuteAsync
    -0.25
     Photocase
    -0.24
     recevrez
    -0.24
    POSITIVE LOGITS
    los
    0.57
    star
    0.56
    cor
    0.56
    port
    0.55
    pot
    0.55
    da
    0.54
    bud
    0.54
    real
    0.54
    lot
    0.54
    lock
    0.54
    Activations Density 0.047%

    No Known Activations