© 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. 21-GEMMASCOPE-TRANSCODER-16K
    7. 1879
    Prev
    Next
    INDEX
    Explanations

    strings of 2-3 letters with a high proportion of the letters `a`, `i`, `o`, `u`, `y`, and/or numbers close together

    oai_token-act-pair · gemini-2.0-flash

    specific code patterns or identifiers

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

    Foreign language/code snippets

    np_max-act-logits · gemini-2.0-flash
    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
    criture
    -0.44
     achter
    -0.44
    insensitive
    -0.43
    isle
    -0.43
    tymologie
    -0.41
    áh
    -0.41
    EYE
    -0.39
    apimachinery
    -0.38
     Sieg
    -0.38
    Tensi
    -0.38
    POSITIVE LOGITS
    windowFixed
    0.74
     surla
    0.72
    الحياه
    0.71
     SwitchCompat
    0.70
     doInBackground
    0.68
     يتيمه
    0.65
    DebuggerNonUser
    0.62
    ់
    0.61
     skak
    0.60
     ThemeData
    0.60
    Activations Density 1.684%

    No Known Activations