© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 2-GEMMASCOPE-TRANSCODER-16K
    7. 13952
    Prev
    Next
    INDEX
    Explanations

    a mix of French and Scandinavian language HTML and Javascript code snippets along with legal terminology

    oai_token-act-pair · gemini-2.0-flash

    other languages

    np_max-act · gemini-2.0-flash

    French Method used: 1 — because top activating tokens are French; explanation: French

    np_max-act-logits · gpt-5-mini
    Triggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_2/width_16k/average_l0_49
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.2.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     Monfieur
    -0.94
     متعلقه
    -0.89
     Majefty
    -0.83
     Theſe
    -0.82
     Económica
    -0.82
     myſelf
    -0.82
     auffi
    -0.81
     виправивши
    -0.81
     himſelf
    -0.79
     ―――――
    -0.79
    POSITIVE LOGITS
    <bos>
    0.82
     de
    0.68
     tra
    0.58
     na
    0.55
     aux
    0.53
    olsom
    0.52
     van
    0.52
    …
    0.52
     PrintWriter
    0.52
    fjord
    0.52
    Activations Density 0.593%

    No Known Activations