© 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. 24-GEMMASCOPE-TRANSCODER-16K
    7. 7620
    Prev
    Next
    INDEX
    Explanations

    a mix of seemingly unrelated tokens including code, bible verses, and abbreviations

    oai_token-act-pair · gemini-2.0-flash

    Code and groups

    np_max-act-logits · gemini-2.0-flash

    say "capital letters" Method used: 2 — tokens after max token start with capital letters; say "capital letters"

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_24/width_16k/average_l0_37
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.24.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
     Μ
    -0.67
    g
    -0.61
     g
    -0.61
     Ма
    -0.60
    ftagPool
    -0.56
    Ｍ
    -0.56
     Me
    -0.56
    Μ
    -0.55
     More
    -0.55
     Ma
    -0.55
    POSITIVE LOGITS
     Савезне
    0.52
     m
    0.52
    Przypisy
    0.50
     Gu
    0.49
     modelAndView
    0.47
     للاسماء
    0.47
     Gi
    0.47
    mær
    0.46
    Gt
    0.46
     Grid
    0.46
    Activations Density 7.008%

    No Known Activations