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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    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. 46578
    Prev
    Next
    INDEX
    Explanations

    closing braces and associated code structure elements in programming

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

    auto-generated code

    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
    sizeCache
    -0.71
    EndInit
    -0.71
    zzleHttp
    -0.68
    MigrationBuilder
    -0.63
    CppMethod
    -0.63
    UnitTesting
    -0.63
    ftagPool
    -0.61
    PerformLayout
    -0.61
     AssemblyTitle
    -0.61
    TagNumber
    -0.60
    POSITIVE LOGITS
    Auto
    0.34
    pu
    0.34
    ∓
    0.33
    #!
    0.33
    <eos>
    0.32
    型
    0.31
    expandindo
    0.31
    PU
    0.31
    ?”
    0.31
    }$​
    0.30
    Activations Density 0.009%

    No Known Activations