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. 1251
    Prev
    Next
    INDEX
    Explanations

    attends to the token "United" from tokens referring to "States."

    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
    0:0.01
    1:0.01
    2:0.55
    3:0.04
    4:0.02
    5:0.01
    6:0.02
    7:0.01
    8:0.00
    9:0.01
    10:0.00
    11:0.00
    12:0.01
    13:0.12
    14:0.09
    15:0.00
    Negative Logits
     directly
    -0.25
     would
    -0.23
     in
    -0.22
    -
    -0.22
     to
    -0.22
     work
    -0.22
     back
    -0.21
     together
    -0.21
    >|</
    -0.21
    ึ้น
    -0.21
    POSITIVE LOGITS
     zwiſchen
    0.44
    OGND
    0.42
    $_['
    0.41
    ஢
    0.41
     ویکی‌پدی
    0.41
     Menſchen
    0.41
     FacebookSign
    0.40
    𑄮
    0.40
    ruik
    0.40
     queſto
    0.40
    Activations Density 0.047%

    No Known Activations