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-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 967
    Prev
    Next
    INDEX
    Explanations

    references to legal documentation or court proceedings

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

    punctuation marks, particularly periods and parentheses

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

    . ( )

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
     They
    -0.90
     These
    -0.86
     There
    -0.84
     We
    -0.80
     It
    -0.77
     However
    -0.77
     Those
    -0.76
     Nevertheless
    -0.75
     This
    -0.75
     Our
    -0.75
    POSITIVE LOGITS
    …~
    0.60
    Референце
    0.58
    下载附件
    0.57
    CppMethod
    0.55
    DockStyle
    0.54
    filepath
    0.52
    LElement
    0.52
    ipheral
    0.51
    URIComponent
    0.51
    englanniksi
    0.51
    Activations Density 0.288%

    No Known Activations