© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 21-GEMMASCOPE-TRANSCODER-16K
    7. 6408
    Prev
    Next
    INDEX
    Explanations

    the word "to", preceeded by a verb, often "have," "has," "is," "are", or "be"

    oai_token-act-pair · gemini-2.0-flash

    variations of the word 'do'

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

    "to do with"

    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
    StructEnd
    -0.70
    fromnode
    -0.68
    Espèce
    -0.66
    новниш
    -0.63
    RTSC
    -0.62
     companyId
    -0.57
    principalColumn
    -0.57
    횟
    -0.57
    例句
    -0.57
    TextBoxColumn
    -0.56
    POSITIVE LOGITS
     do
    2.55
    do
    2.14
     Do
    1.89
    Do
    1.86
     DO
    1.63
     does
    1.41
    DO
    1.38
     did
    1.21
     doing
    1.18
     doe
    1.10
    Activations Density 1.184%

    No Known Activations