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. 14171
    Prev
    Next
    INDEX
    Explanations

    commands or prompts for user interactions in a technical context

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

    Formatting or code-related tokens

    np_acts-logits-general · gemini-2.0-flash

    structured lists and annotations

    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
    <eos>
    -0.66
    onAttach
    -0.58
    modelBuilder
    -0.56
     архивлан
    -0.53
    bootstrapcdn
    -0.50
    TestingModule
    -0.49
    "?>
    -0.48
    DropTable
    -0.48
     safely
    -0.47
    loth
    -0.47
    POSITIVE LOGITS
    IntoConstraints
    0.66
    NewUrlParser
    0.61
    Jîn
    0.58
     calendriers
    0.55
     magasiner
    0.55
    κτηρισ
    0.53
     tille
    0.52
    Carriera
    0.52
    getSelectedItem
    0.52
    曖昧さ回避
    0.52
    Activations Density 0.025%

    No Known Activations