© 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. 23-GEMMASCOPE-TRANSCODER-16K
    7. 11854
    Prev
    Next
    INDEX
    Explanations

    numbers, especially in legal or business documents

    oai_token-act-pair · gemini-2.0-flash

    legal and technical documents

    np_max-act-logits · gemini-2.0-flash

    ▁ Method used: 1 — because MAX_ACTIVATING_TOKENS are identical

    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_23/width_16k/average_l0_25
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.23.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
    <
    -0.65
    \\
    -0.64
    \
    -0.62
    “
    -0.62
    A
    -0.57
    &
    -0.56
    The
    -0.56
    *
    -0.55
    //
    -0.53
    —
    -0.52
    POSITIVE LOGITS
     resourceCulture
    1.05
    tvguidetime
    1.04
    IVEREF
    1.01
     AssemblyCulture
    1.00
    protoimpl
    0.99
    SBATCH
    0.96
    VIAF
    0.94
    NameInMap
    0.94
    esgue
    0.93
     Normdatei
    0.91
    Activations Density 3.090%

    No Known Activations