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-2-9B
    4. Attention Out - 16k
    5. 0-GEMMASCOPE-ATT-16K
    6. 1312
    Prev
    Next
    INDEX
    Explanations

    attends to the word "ultimately" from various later tokens that introduce a pause or punctuation

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61
    Prompts (Dashboard)
    16,384 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
    1:0.01
    2:0.11
    3:0.02
    4:0.00
    5:0.01
    6:0.10
    7:0.08
    8:0.01
    9:0.01
    10:0.01
    11:0.01
    12:0.02
    13:0.34
    14:0.16
    15:0.02
    Negative Logits
     zijne
    -0.72
     Italij
    -0.71
     miniaturka
    -0.71
     aislado
    -0.69
     desmotivaciones
    -0.69
     muerta
    -0.69
     herida
    -0.69
     vuestros
    -0.69
     vuestro
    -0.69
     flotante
    -0.68
    POSITIVE LOGITS
     Fac
    0.67
     Box
    0.66
     Sc
    0.66
     cad
    0.66
     Bra
    0.66
     Tri
    0.65
     gre
    0.65
     Dia
    0.64
     Pen
    0.64
     Str
    0.64
    Activations Density 0.096%

    No Known Activations