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. 0-GEMMASCOPE-ATT-16K
    6. 64
    Prev
    Next
    INDEX
    Explanations

    attends to championship-related tokens from non-championship related tokens

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k
    Configuration
    google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104
    Prompts (Dashboard)
    36,864 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
    1:0.02
    2:0.05
    3:0.60
    4:0.06
    5:0.03
    6:0.11
    7:0.08
    Negative Logits
     myſelf
    -0.48
     Efq
    -0.44
     pngtree
    -0.43
     doubtnut
    -0.41
     Monfieur
    -0.41
     vPvB
    -0.41
     Chriftian
    -0.40
    
    -0.40
     Shakspeare
    -0.40
    دانشنامهٔ
    -0.40
    POSITIVE LOGITS
     imp
    0.26
     his
    0.25
     contra
    0.25
     or
    0.25
     r
    0.24
    過ぎる
    0.24
     '
    0.24
     R
    0.24
     for
    0.24
     видите
    0.23
    Activations Density 2.472%

    No Known Activations