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. 10-GEMMASCOPE-ATT-16K
    6. 5967
    Prev
    Next
    INDEX
    Explanations

    attends to specific tokens representing numeric values from a variety of temporal or contextual elements represented by brackets

    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
    0:0.13
    1:0.19
    2:0.19
    3:0.08
    4:0.11
    5:0.08
    6:0.05
    7:0.13
    Negative Logits
     uParam
    -0.27
    minecraft
    -0.25
    vably
    -0.24
    goog
    -0.24
    adaş
    -0.24
    +:+
    -0.23
     Incorporated
    -0.23
     مشين
    -0.22
    badlogic
    -0.22
    stalgia
    -0.22
    POSITIVE LOGITS
    Geplaatst
    0.40
     avoient
    0.33
    UnknownFields
    0.32
     lenker
    0.32
    AddTagHelper
    0.31
     beginnetje
    0.30
     auroit
    0.30
    ┛
    0.30
    UnsafeEnabled
    0.30
     feroit
    0.29
    Activations Density 0.128%

    No Known Activations