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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 21-GEMMASCOPE-TRANSCODER-16K
    7. 4085
    Prev
    Next
    INDEX
    Explanations

    words in French

    oai_token-act-pair · gemini-2.0-flash

    technical terms related to programming or software development

    np_token-act-pair-logits · gpt-4o-mini

    Non-English language text

    np_max-act-logits · gemini-2.0-flash

    French articles/possessives Method used: 1. Reason: tokens like "▁mon, ▁un, ▁leurs, ▁les, ▁des, ▁son, ▁le" are French articles/possessives

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     betweenstory
    -0.67
    ysuckle
    -0.61
     lids
    -0.57
     individuale
    -0.56
    lider
    -0.53
     Woodford
    -0.53
    uxxxx
    -0.52
     Questi
    -0.52
     souvenir
    -0.51
     farine
    -0.51
    POSITIVE LOGITS
    SerializedSize
    0.61
    ArgsConstructor
    0.54
     AssemblyProduct
    0.52
    InstrumentedTest
    0.50
    
    0.47
    windowFixed
    0.47
    Vanjske
    0.45
    calculus
    0.45
     jScrollPane
    0.45
    IMDG
    0.44
    Activations Density 0.342%

    No Known Activations