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. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13755
    Prev
    Next
    INDEX
    Explanations

    words related to actions, especially those indicating instigation or involvement in events

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    ending in "ate", "ated", "ating", "ation", or "ations"

    np_acts-logits-general · gemini-2.0-flash

    predominates, instigates, negotiates

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     AppComponent
    -0.67
    LayoutStyle
    -0.56
     utafitiHapana
    -0.54
    Datuak
    -0.53
     NgModule
    -0.52
    ojo
    -0.49
    gráfica
    -0.49
    ########.
    -0.49
     soste
    -0.48
    excellents
    -0.48
    POSITIVE LOGITS
    ated
    1.87
    ating
    1.81
    ation
    1.62
    ATED
    1.54
    ations
    1.45
    ates
    1.37
    ATING
    1.32
    ATION
    1.31
    ate
    1.30
    ators
    1.30
    Activations Density 0.492%

    No Known Activations