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-2-9B
    4. Attention Out - 16k
    5. 0-GEMMASCOPE-ATT-16K
    6. 1236
    Prev
    Next
    INDEX
    Explanations

    attends to the token "avoid" from various tokens related to negative outcomes or issues

    oai_attention-head · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61
    Prompts (Dashboard)
    16,384 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
    0:0.01
    1:0.01
    2:0.11
    3:0.01
    4:0.01
    5:0.01
    6:0.32
    7:0.16
    8:0.00
    9:0.01
    10:0.00
    11:0.01
    12:0.02
    13:0.10
    14:0.14
    15:0.01
    Negative Logits
     Bewußt
    -0.39
     prohibido
    -0.38
    kháu
    -0.37
     betweenstory
    -0.36
     productivo
    -0.36
     lección
    -0.36
     pierna
    -0.35
     benih
    -0.35
     Pferd
    -0.35
     autorytatywna
    -0.35
    POSITIVE LOGITS
    $
    0.30
    IBOutlet
    0.30
    ResumeLayout
    0.29
    BP
    0.29
    	
    0.28
    GroupLayout
    0.27
    arg
    0.27
    HeaderText
    0.27
    toThrow
    0.26
    setcounter
    0.26
    Activations Density 0.059%

    No Known Activations