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. 6102
    Prev
    Next
    INDEX
    Explanations

    attends to possessive tokens from equivalent noun-related tokens

    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
    0:0.23
    1:0.21
    2:0.27
    3:0.05
    4:0.04
    5:0.02
    6:0.03
    7:0.11
    Negative Logits
     يتيمه
    -0.45
    BeginContext
    -0.39
    mybatisplus
    -0.38
    aarrggbb
    -0.38
    contentLoaded
    -0.38
     ویکی‌پدیا
    -0.37
     numberWith
    -0.36
     ModelExpression
    -0.36
     }:
    -0.34
    Становништво
    -0.34
    POSITIVE LOGITS
    Su
    0.23
    becca
    0.23
    dafx
    0.22
     su
    0.22
    laufen
    0.21
    lloworld
    0.21
    wara
    0.20
     súčas
    0.20
    lihatan
    0.20
    minlength
    0.20
    Activations Density 0.445%

    No Known Activations