Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 6952
    Prev
    Next
    INDEX
    Explanations

    attends to the closure of brackets and similar constructs from the positions of tokens either preceding or following them

    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
    0:0.07
    1:0.09
    2:0.27
    3:0.08
    4:0.10
    5:0.05
    6:0.14
    7:0.16
    Negative Logits
     للمعارف
    -0.39
     IndexPath
    -0.36
     createSlice
    -0.34
    UnusedPrivate
    -0.33
     smtplib
    -0.31
     BoxFit
    -0.30
    sizeCache
    -0.30
    PostExecute
    -0.30
    Istorija
    -0.29
    borderSide
    -0.29
    POSITIVE LOGITS
    AndEndTag
    0.32
     tartalomajánló
    0.30
    ter
    0.29
    brot
    0.29
    tern
    0.28
     !-
    0.28
    eare
    0.28
    tore
    0.28
    InputModule
    0.27
    caris
    0.27
    Activations Density 0.689%

    No Known Activations