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. 10-GEMMASCOPE-ATT-16K
    6. 6070
    Prev
    Next
    INDEX
    Explanations

    attends to the token "amsmath" from the repeated instances of "ams" used in a LaTeX context

    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
    0:0.11
    1:0.20
    2:0.10
    3:0.07
    4:0.22
    5:0.03
    6:0.12
    7:0.11
    Negative Logits
    ftagPool
    -0.48
    gridx
    -0.45
    complexContent
    -0.38
    GEBURTSDATUM
    -0.35
    GrantedAuthority
    -0.35
     competir
    -0.34
    harusnya
    -0.34
    javac
    -0.34
    forName
    -0.33
    GeneratedCode
    -0.33
    POSITIVE LOGITS
    ?";
    0.31
    $/,
    0.31
     $\}$
    0.31
    =";
    0.30
    ();)
    0.30
    ;";
    0.30
    TagHelper
    0.30
    ;';
    0.29
     sugg
    0.29
     ſche
    0.29
    Activations Density 0.008%

    No Known Activations