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-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 91279
    Prev
    Next
    INDEX
    Explanations

    references to popes and the Catholic Church

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

    Pope and Vatican

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    romyalgia
    -0.63
     Saxons
    -0.61
    classnames
    -0.60
    tershire
    -0.58
    imental
    -0.58
     Euripides
    -0.57
    PreferredItem
    -0.57
    rettes
    -0.57
    odils
    -0.57
    はじめに
    -0.56
    POSITIVE LOGITS
     Vatican
    0.68
     Vaticano
    0.65
     Relaciones
    0.52
     institución
    0.51
     pestaña
    0.47
     Мексичка
    0.45
     Pope
    0.45
     Sosten
    0.42
     Papst
    0.42
     âgées
    0.41
    Activations Density 0.008%

    No Known Activations