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. 9148
    Prev
    Next
    INDEX
    Explanations

    specific references to organizations, events, and titles related to healthcare and technology

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

    Follows punctuation or symbols

    np_acts-logits-general · gemini-2.0-flash

    snapshot, gene therapy, F1, 2017

    np_acts-logits-general · gemini-2.5-flash-lite

    typographic and formatting markers—especially apostrophes, ampersands, hyphenated terms, punctuated abbreviations, and year-like numerals.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
     ویکی‌پدی
    -0.49
     Roskov
    -0.47
    انجليز
    -0.45
     Мексичка
    -0.43
    UnusedPrivate
    -0.41
     disambiguazione
    -0.41
     Taktlose
    -0.41
     ویکی‌پدیای
    -0.39
     MwSt
    -0.39
    verifyException
    -0.36
    POSITIVE LOGITS
    AddField
    0.48
    orship
    0.47
    Plo
    0.47
    vint
    0.46
     iprot
    0.46
    beep
    0.45
     Ety
    0.45
    plo
    0.45
     Fill
    0.45
    FontStyle
    0.44
    Activations Density 4.464%

    No Known Activations