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. 5923
    Prev
    Next
    INDEX
    Explanations

    quotation marks and their context in relation to other words

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

    backend problem or hell is going

    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
     للمعارف
    -0.75
    fromnode
    -0.68
    LookAnd
    -0.61
    genodigd
    -0.60
     betweenstory
    -0.58
     Wicidata
    -0.57
     BorderSide
    -0.56
     otomatig
    -0.56
    UrlResolution
    -0.54
     noDo
    -0.54
    POSITIVE LOGITS
     StatefulWidget
    0.40
    fixture
    0.39
    ClientRect
    0.36
    Arg
    0.36
    ेश
    0.35
     ***
    0.35
     ch
    0.35
    MENAFN
    0.34
     continuation
    0.34
    bout
    0.34
    Activations Density 0.013%

    No Known Activations