Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 21-GEMMASCOPE-RES-16K
    6. 12521
    Prev
    Next
    INDEX
    Explanations

    elements related to markup or document structure conventions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 21-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_21/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.21.hook_resid_post
    Hook Layer
    21
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     the
    -0.56
     sta
    -0.49
     of
    -0.43
     ter
    -0.43
     per
    -0.42
     a
    -0.42
     du
    -0.41
     g
    -0.40
    ERSHIP
    -0.40
    pola
    -0.40
    POSITIVE LOGITS
     autorytatywna
    0.97
     تضيفلها
    0.92
    RegressionTest
    0.92
    GEBURTSDATUM
    0.81
     pleaſure
    0.79
    Coordonnées
    0.78
    AutoScale
    0.78
     îna
    0.77
    évaluateur
    0.73
    ſelves
    0.72
    Activations Density 0.663%

    No Known Activations