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. 6432
    Prev
    Next
    INDEX
    Explanations

    attends to tokens that have "der" from tokens later that include "der."

    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
    0:0.05
    1:0.16
    2:0.05
    3:0.03
    4:0.11
    5:0.48
    6:0.03
    7:0.05
    Negative Logits
    RectangleBorder
    -0.45
    verwijspagina
    -0.33
    ngdoc
    -0.32
     FetchType
    -0.32
    BagConstraints
    -0.31
    InputBorder
    -0.31
    chafter
    -0.31
    nonatomic
    -0.30
    UrlResolution
    -0.30
    EventLoop
    -0.30
    POSITIVE LOGITS
     podjela
    0.32
    帖最后由
    0.31
    GGLE
    0.30
     Савезне
    0.30
    chlag
    0.27
    edly
    0.27
    LookAnd
    0.27
    nessed
    0.27
    AndWait
    0.27
    atamente
    0.27
    Activations Density 0.064%

    No Known Activations