© 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-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 71964
    Prev
    Next
    INDEX
    Explanations

    patterns of whitespace or breaks in text

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

    code snippets C/C++/JS/Python

    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
     Paglinawan
    -0.65
    EndContext
    -0.60
     estekak
    -0.58
    новниш
    -0.56
     Administrativna
    -0.53
    tagHelperRunner
    -0.51
    ValueStyle
    -0.51
    CppCodeGen
    -0.50
    Rhestr
    -0.49
    gestone
    -0.48
    POSITIVE LOGITS
    entras
    0.39
     Autorizaciones
    0.37
    Lé
    0.37
     actionMode
    0.34
    éndolo
    0.33
     '+':
    0.33
     maneiras
    0.33
    いいですね
    0.33
    Климат
    0.33
    näm
    0.32
    Activations Density 0.018%

    No Known Activations