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. 374
    Prev
    Next
    INDEX
    Explanations

    attends to potential or conditional scenarios indicated by "could" from tokens that provide context or circumstances that allow for this possibility

    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
    0:0.41
    1:0.01
    2:0.03
    3:0.01
    4:0.10
    5:0.05
    6:0.05
    7:0.30
    Negative Logits
     be
    -0.81
     do
    -0.60
     have
    -0.59
     get
    -0.58
     see
    -0.57
     the
    -0.56
     are
    -0.52
     (
    -0.51
    .
    -0.51
     can
    -0.50
    POSITIVE LOGITS
     myſelf
    1.03
     Theſe
    1.03
     itſelf
    1.02
     ſeveral
    1.00
     Monfieur
    0.99
     Efq
    0.98
     pleaſure
    0.96
     purpoſe
    0.96
     Houſe
    0.94
     ſever
    0.94
    Activations Density 0.055%

    No Known Activations