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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 14523
    Prev
    Next
    INDEX
    Explanations

    various aspects related to legal and regulatory contexts

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

    Preceding time-related words (current, parents, warnings, etc.)

    np_acts-logits-general · gemini-2.0-flash

    temporal correlations

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 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
    UserScript
    -0.77
     nakalista
    -0.64
    ViewFeatures
    -0.58
    Искәрмәләр
    -0.57
     Vikipedi
    -0.56
     ویکی‌پدی
    -0.54
    Autoritní
    -0.54
    хьтан
    -0.52
    
    -0.51
    Rohy
    -0.49
    POSITIVE LOGITS
     when
    0.53
     during
    0.49
     Gegenteil
    0.40
     after
    0.40
     anzunehmen
    0.39
     atunci
    0.39
     cuando
    0.39
    เพื่อ
    0.38
     kiedy
    0.37
     whenever
    0.36
    Activations Density 0.753%

    No Known Activations