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. Residual Stream - 16k
    5. 2-GEMMASCOPE-RES-16K
    6. 1195
    Prev
    Next
    INDEX
    Explanations

    instances of visual attention and observation, particularly focusing on the action of looking

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 2-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_2/width_16k/average_l0_141
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    뀜
    -0.80
     fetus
    -0.67
     sessionFactory
    -0.67
    뀐
    -0.64
    BackStack
    -0.63
     lī
    -0.62
    êmica
    -0.62
    servez
    -0.61
     coscienza
    -0.60
     idéia
    -0.60
    POSITIVE LOGITS
     looked
    2.52
    looked
    2.08
     Looked
    2.03
     looks
    1.83
     Looks
    1.79
    Looks
    1.74
     looking
    1.65
    looks
    1.55
     Looking
    1.53
    Looking
    1.52
    Activations Density 0.094%

    No Known Activations