© 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. 22-GEMMASCOPE-TRANSCODER-16K
    7. 15915
    Prev
    Next
    INDEX
    Explanations

    code fragments, probably C++ or Java

    oai_token-act-pair · gemini-2.0-flash

    say "="

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_22/width_16k/average_l0_15
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.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
    ChildScrollView
    -1.07
    VersionUID
    -0.96
    IndentedString
    -0.94
    </tfoot>
    -0.90
    enumii
    -0.88
    ngdoc
    -0.88
    WriteBarrier
    -0.88
    migrationBuilder
    -0.85
     transfieras
    -0.85
    BeginContext
    -0.84
    POSITIVE LOGITS
    =
    0.66
     blé
    0.46
    per
    0.44
    (
    0.44
     löyty
    0.41
     kuulu
    0.41
     kuten
    0.39
     uova
    0.39
     tapaht
    0.38
     kautta
    0.38
    Activations Density 6.425%

    No Known Activations