© 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. 3082
    Prev
    Next
    INDEX
    Explanations

    the labels of wikipedia categories, with some activation on past-tense verbs such as 'based' or 'directed'

    oai_token-act-pair · gemini-2.0-flash

    categories related to politics and films

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

    Category

    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
     Theſe
    -1.84
     itſelf
    -1.70
     Мексичка
    -1.69
     Monfieur
    -1.62
     myſelf
    -1.61
     Efq
    -1.53
     ModelExpression
    -1.50
     purpoſe
    -1.44
    parsedMessage
    -1.44
    <unused43>
    -1.42
    POSITIVE LOGITS
     
    0.90
    .
    0.85
    ↵↵
    0.85
    ,
    0.84
    2
    0.84
    </strong>
    0.83
    1
    0.82
    3
    0.79
    ’
    0.79
     de
    0.77
    Activations Density 0.263%

    No Known Activations