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. 5600
    Prev
    Next
    INDEX
    Explanations

    attends to conjunctions such as "and" from subsequent tokens indicating a transition to related or dependent actions

    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
    0:0.11
    1:0.13
    2:0.16
    3:0.15
    4:0.11
    5:0.05
    6:0.11
    7:0.14
    Negative Logits
    
    -0.27
     []:
    -0.26
     bille
    -0.23
    öld
    -0.23
    agrama
    -0.22
    amkeit
    -0.22
     riuscito
    -0.22
     للمعارف
    -0.22
     know
    -0.22
     برانيه
    -0.22
    POSITIVE LOGITS
    AddTagHelper
    0.33
     Efq
    0.32
     HasFactory
    0.31
    fjspx
    0.28
     useCallback
    0.25
     AssemblyCompany
    0.25
     unknownFields
    0.24
    ExtendWith
    0.24
    EndContext
    0.23
     Normdatei
    0.23
    Activations Density 0.104%

    No Known Activations