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. 5799
    Prev
    Next
    INDEX
    Explanations

    attends to numerical tokens from the related punctuation 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
    0:0.06
    1:0.13
    2:0.06
    3:0.05
    4:0.21
    5:0.30
    6:0.06
    7:0.09
    Negative Logits
    Personendaten
    -0.30
    OOTDTY
    -0.29
    AutoScale
    -0.29
     AssemblyProduct
    -0.29
    Дереккөздер
    -0.28
     CURLOPT
    -0.28
     &___
    -0.28
    $_['
    -0.28
    upakan
    -0.28
    Numerology
    -0.28
    POSITIVE LOGITS
     sibi
    0.29
     spira
    0.26
     quæ
    0.26
    Diweddarwch
    0.25
    脚注の使い方
    0.22
    ktive
    0.22
    もっと
    0.21
    setVerticalGroup
    0.20
    ResponseWriter
    0.20
     omnia
    0.20
    Activations Density 0.245%

    No Known Activations