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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 12731
    Prev
    Next
    INDEX
    Explanations

    keywords related to class and method definitions in programming languages

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

    Code or numbers followed by punctuation

    np_acts-logits-general · gemini-2.0-flash

    programming language keywords and punctuation

    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-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    self
    -0.52
    te
    -0.48
    irdre
    -0.45
    ($("#
    -0.42
     needs
    -0.41
     współ
    -0.39
    '
    -0.39
    ủ
    -0.39
    <bos>
    -0.38
     self
    -0.38
    POSITIVE LOGITS
     boolean
    1.22
     bool
    1.09
    boolean
    1.07
     Boolean
    1.06
    bool
    0.99
     int
    0.97
     String
    0.97
    Boolean
    0.95
     float
    0.90
     double
    0.89
    Activations Density 0.145%

    No Known Activations