© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 542
    Prev
    Next
    INDEX
    Explanations

    sequences of integers and numeric patterns

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

    Hexadecimal representation

    np_acts-logits-general · gemini-2.0-flash

    hex numbers and code syntax

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
     îna
    -0.54
    webElementXpaths
    -0.53
    Personendaten
    -0.51
    archiwizowane
    -0.50
     AssemblyVersion
    -0.49
    SharedCtor
    -0.47
    tagPool
    -0.47
     Geſch
    -0.47
    fjspx
    -0.46
    pecabe
    -0.46
    POSITIVE LOGITS
    ViewImports
    0.60
     @"/
    0.56
     wireType
    0.38
     defaultstate
    0.33
     subs
    0.32
    MessageOf
    0.32
     porto
    0.32
     intermedia
    0.32
     ſta
    0.30
    httphttps
    0.29
    Activations Density 0.646%

    No Known Activations