© 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-2-2B
    7. Residual Stream - 65k
    9. 20-GEMMASCOPE-RES-65K
    11. 543
    Prev
    Next
    INDEX
    Explanations

    code structures and syntax elements in programming

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

    closing braces and indentation-related syntax in code

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

    procedural protections, mixing results

    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
     BRO
    -0.44
    ubel
    -0.43
     Bo
    -0.42
    нос
    -0.42
     Araujo
    -0.42
     sebe
    -0.41
    éről
    -0.41
    bro
    -0.41
    rig
    -0.41
    幸
    -0.41
    POSITIVE LOGITS
     فريبيس
    1.00
     StatefulWidget
    0.99
     متعلقه
    0.97
    OGND
    0.93
    adaptiveStyles
    0.89
     AssemblyVersion
    0.89
    expandindo
    0.88
     useAppContext
    0.86
    tagHelperRunner
    0.86
    MigrationBuilder
    0.85
    Activations Density 0.021%

    No Known Activations