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. 5715
    Prev
    Next
    INDEX
    Explanations

    attends to the token "and" from tokens that are connected or grouped in some way

    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
    0:0.12
    1:0.11
    2:0.12
    3:0.07
    4:0.07
    5:0.03
    6:0.09
    7:0.34
    Negative Logits
     Roskov
    -0.28
    :✨
    -0.28
    üyada
    -0.25
     scolaires
    -0.25
    Kaynakça
    -0.25
     naturelles
    -0.24
    kháu
    -0.24
     mussel
    -0.24
    Ранее
    -0.23
    xtext
    -0.23
    POSITIVE LOGITS
    endwhile
    0.36
    RenderAtEndOf
    0.36
    Datuak
    0.31
    AndEndTag
    0.31
    TemporalType
    0.30
    MENAFN
    0.29
    phazard
    0.29
    存于互联网档案馆
    0.28
    audiovisuel
    0.27
     ostavi
    0.26
    Activations Density 0.145%

    No Known Activations