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. 1255
    Prev
    Next
    INDEX
    Explanations

    attends to incorrect response tokens from correct response tokens

    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
    0:0.13
    1:0.02
    2:0.01
    3:0.03
    4:0.04
    5:0.04
    6:0.08
    7:0.08
    8:0.06
    9:0.07
    10:0.03
    11:0.05
    12:0.08
    13:0.11
    14:0.04
    15:0.05
    Negative Logits
    <bos>
    -0.20
     the
    -0.18
     step
    -0.15
     surface
    -0.14
     side
    -0.12
     distance
    -0.12
     sides
    -0.11
     square
    -0.10
     narrow
    -0.10
     lower
    -0.10
    POSITIVE LOGITS
    <unused99>
    1.12
    JAKARTA
    1.05
     ſoll
    1.00
    FIFA
    0.98
     ſua
    0.97
    
    0.96
     églises
    0.96
     ſeinem
    0.95
     coachTry
    0.95
     royaume
    0.95
    Activations Density 0.016%

    No Known Activations