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. 150
    Prev
    Next
    INDEX
    Explanations

    attends to the token "skill" from related tokens "set"

    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
    0:0.54
    1:0.02
    2:0.03
    3:0.03
    4:0.19
    5:0.09
    6:0.02
    7:0.04
    Negative Logits
     Monfieur
    -0.67
     Majefty
    -0.60
     Houſe
    -0.59
     Reſ
    -0.56
     Diſ
    -0.56
     Anſ
    -0.55
     Shakspeare
    -0.54
     Inſ
    -0.53
     Forumite
    -0.52
     purpoſe
    -0.52
    POSITIVE LOGITS
     —
    0.33
    <eos>
    0.33
     T
    0.33
     O
    0.30
     D
    0.29
     N
    0.29
     C
    0.29
     P
    0.29
     at
    0.28
     but
    0.28
    Activations Density 0.017%

    No Known Activations