© 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. 8-GEMMASCOPE-TRANSCODER-16K
    7. 4002
    Prev
    Next
    INDEX
    Explanations

    references to sporting events and competition, including winning and losing records

    oai_token-act-pair · gemini-2.0-flash

    sports, winning Method used: 2, reason "sports, winning" is more accurate than just "victory", explanation: sports, winning

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_8/width_16k/average_l0_52
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.8.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
    ajs
    -0.42
    getInteger
    -0.40
     &
    -0.40
     ->
    -0.39
     →
    -0.38
     fino
    -0.37
    <bos>
    -0.36
    MAA
    -0.35
    limus
    -0.35
    hide
    -0.35
    POSITIVE LOGITS
     Monfieur
    0.89
     purpoſe
    0.78
     Houſe
    0.77
    ſelf
    0.74
    JspWriter
    0.74
     Efq
    0.73
     NSCoder
    0.73
     houſe
    0.73
     quæ
    0.71
     Theſe
    0.69
    Activations Density 1.291%

    No Known Activations