© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 592
    Prev
    Next
    INDEX
    Explanations

    references to data types and properties in JSON or similar structured formats

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

    JSON or structured data syntax elements, particularly type-related tokens.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 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
    endphp
    -0.69
    UnitTesting
    -0.57
     bénévol
    -0.56
    postsleuth
    -0.54
     initWith
    -0.54
    skjaer
    -0.52
    archiviato
    -0.52
    RUnlock
    -0.51
    !("{}",
    -0.51
     gewä
    -0.51
    POSITIVE LOGITS
     type
    1.37
     Type
    1.36
    Type
    1.29
     TYPE
    1.28
    type
    1.22
    TYPE
    1.17
     types
    1.08
    getType
    1.08
     getType
    1.07
    etype
    1.05
    Activations Density 0.232%

    No Known Activations