© 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. 2-GEMMASCOPE-TRANSCODER-16K
    7. 10060
    Prev
    Next
    INDEX
    Explanations

    conjunctions, especially "and", and some related words like "where", "oh", "step", "ra", and "soever"

    oai_token-act-pair · gemini-2.0-flash

    and

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_2/width_16k/average_l0_49
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.2.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
     myſelf
    -1.30
     whoſe
    -1.28
     himſelf
    -1.28
     itſelf
    -1.27
     Monfieur
    -1.24
     ſeveral
    -1.23
     themſelves
    -1.23
     Diſ
    -1.22
     Efq
    -1.20
     becauſe
    -1.17
    POSITIVE LOGITS
    WriteTagHelper
    0.55
    chures
    0.52
     ModelExpression
    0.48
     Reverso
    0.48
     AlertDialog
    0.46
    Tembelea
    0.45
     onAnimation
    0.45
     Walkover
    0.43
    wikidata
    0.43
    Slight
    0.42
    Activations Density 0.576%

    No Known Activations