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-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 2418
    Prev
    Next
    INDEX
    Explanations

    references to silicon dioxide (Si) in various contexts

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

    DSi concentration and transport

    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
     duele
    -0.39
    挑戦
    -0.38
     Camaro
    -0.37
     tiras
    -0.36
     iniciado
    -0.36
     mennes
    -0.36
    ibut
    -0.35
     inici
    -0.35
     manusia
    -0.35
     industriels
    -0.35
    POSITIVE LOGITS
    Nintendo
    0.81
     Nintendo
    0.80
     nintendo
    0.72
    nintendo
    0.69
    iibo
    0.68
     Houſe
    0.64
     Efq
    0.59
     Wii
    0.57
     ब्रेकडाउन
    0.55
     Bowser
    0.54
    Activations Density 0.009%

    No Known Activations