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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-2B
    3. 25-GEMMASCOPE-TRANSCODER-16K
    4. 4062
    Prev
    Next
    INDEX
    Explanations

    abbreviations and acronyms, often in all caps

    oai_token-act-pair · gemini-2.0-flash

    Acronyms for Swaziland

    np_max-act-logits · gemini-2.0-flash

    say "two-letter uppercase" Method used: 2 — because the following tokens are all two-letter uppercase sequences — say "two-letter uppercase"

    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_25/width_16k/average_l0_41
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.25.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
     CURLOPT
    -0.35
    em
    -0.32
    cu
    -0.30
    cl
    -0.29
    de
    -0.29
    dd
    -0.28
    gl
    -0.28
    cp
    -0.28
    cm
    -0.27
     ttk
    -0.27
    POSITIVE LOGITS
     purpoſe
    0.94
     pleaſure
    0.93
     houſe
    0.90
     Monfieur
    0.90
     ſtate
    0.89
     myſelf
    0.85
     ſmall
    0.85
     ſtand
    0.85
     themſelves
    0.83
     diſt
    0.82
    Activations Density 18.424%

    No Known Activations