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. 432
    Prev
    Next
    INDEX
    Explanations

    attends to the token "va" from tokens later in the sequence that contain variations of "va" within them

    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
    0:0.03
    1:0.10
    2:0.57
    3:0.05
    4:0.05
    5:0.04
    6:0.07
    7:0.05
    Negative Logits
     itſelf
    -1.86
     myſelf
    -1.85
     Monfieur
    -1.80
     Efq
    -1.73
     Majefty
    -1.66
     purpoſe
    -1.66
     Anſ
    -1.66
     houſe
    -1.64
     Jefus
    -1.63
     pleaſure
    -1.63
    POSITIVE LOGITS
     
    0.96
     (
    0.86
     “
    0.84
     I
    0.82
     "
    0.75
     a
    0.74
     R
    0.73
     in
    0.73
     H
    0.73
     T
    0.73
    Activations Density 0.030%

    No Known Activations