© 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. 9723
    Prev
    Next
    INDEX
    Explanations

    a combination of imperative sentences, emotional punctuation, and words about seeing or being in the know.

    oai_token-act-pair · gemini-2.0-flash

    variations of the word "keep"

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

    keep/eye

    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
    DispatchToProps
    -0.81
    ViewImports
    -0.79
     الحره
    -0.76
    ParallelGroup
    -0.72
    engertian
    -0.69
    Билгалдахарш
    -0.66
     NSCoder
    -0.64
    ⎦
    -0.64
    ')";
    -0.60
    qrstuvwxyz
    -0.59
    POSITIVE LOGITS
     keep
    2.03
     Keep
    1.90
    Keep
    1.85
    keep
    1.79
     keeping
    1.77
     keeps
    1.73
     kept
    1.72
     KEEP
    1.64
     Keeping
    1.59
    KEEP
    1.59
    Activations Density 0.750%

    No Known Activations