© 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. 9505
    Prev
    Next
    INDEX
    Explanations

    hyphenated numbers followed by place names

    oai_token-act-pair · gemini-2.0-flash

    programming-related terms

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

    Documentation and locations

    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
     nadar
    -0.52
     createContext
    -0.51
     >=",
    -0.50
    genomen
    -0.48
     }],
    -0.48
    mögens
    -0.48
     sumpay
    -0.47
    }`).
    -0.47
     erfaren
    -0.47
     nakalista
    -0.47
    POSITIVE LOGITS
    TagMode
    0.62
    ToScroll
    0.58
     CreateTagHelper
    0.57
    __(/*!
    0.54
    atorship
    0.52
    elemField
    0.52
    ndra
    0.52
    transQ
    0.51
    ophil
    0.51
    новниш
    0.50
    Activations Density 1.035%

    No Known Activations