© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 1070
    Prev
    Next
    INDEX
    Explanations

    mathematical terminology and symbols related to rates, measurements, and definitions in scientific contexts

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

    mathematical notation and equations, particularly those involving LaTeX or symbolic mathematical expressions.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    This neuron detects code and programming syntax, particularly focusing on special characters, formatting, and code structure across various programming languages.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    sections of highly technical, structured formatting—code, markup, and mathematical notation—indicating programmatic or academic content blocks.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    code and technical documentation, particularly programming syntax, mathematical notation, and structured data formats.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    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
     Савезне
    -1.23
     Мексичка
    -1.09
    tagHelperRunner
    -1.08
     ویکی‌پدی
    -0.99
    TagMode
    -0.93
    AddTagHelper
    -0.93
    SourceChecksum
    -0.93
    GEBURTSDATUM
    -0.91
     мәкал
    -0.91
    Personensuche
    -0.91
    POSITIVE LOGITS
    0.54
    mathrm
    0.52
    ’
    0.52
    '
    0.48
    begin
    0.43
     culte
    0.36
    hole
    0.34
    obs
    0.33
     
    0.33
     Quest
    0.33
    Activations Density 99.047%
    No Known Activations