© 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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 138
    Prev
    Next
    INDEX
    Explanations

    patterns related to regular expressions and parsing code

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

    Regular expressions and pattern matching

    np_acts-logits-general · gemini-2.0-flash

    regex pattern matching

    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
    •Cluster 31

    No Comments

    Negative Logits
    BeginInit
    -0.89
    исленность
    -0.67
     <>",
    -0.65
    WireFormatLite
    -0.64
    typeparam
    -0.60
     elevator
    -0.58
    enumi
    -0.56
     iVar
    -0.56
     beri
    -0.56
     (!__
    -0.54
    POSITIVE LOGITS
     pattern
    1.41
     Pattern
    1.37
    Pattern
    1.35
     match
    1.27
    pattern
    1.27
     patterns
    1.24
     Match
    1.22
     regex
    1.21
     matches
    1.18
    Match
    1.17
    Activations Density 0.399%

    No Known Activations