© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 50546
    Prev
    Next
    INDEX
    Explanations

    boxing matches and boxers

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    adamkarvonen/qwen3-32b-saes/saes_Qwen_Qwen3-32B_batch_top_k/resid_post_layer_32
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    .sess
    -0.10
     lak
    -0.10
    ellen
    -0.09
    æ¸Ĭ
    -0.09
    icious
    -0.09
    lotte
    -0.09
    isch
    -0.09
    社ä¼ļåĮĸ
    -0.09
    å·½
    -0.09
    累累
    -0.09
    POSITIVE LOGITS
     Boxing
    0.12
    æĭ³
    0.12
     boxing
    0.12
     boxer
    0.10
    æľīäºĨ
    0.09
    ä¼ı
    0.09
    æĭ³å¤´
    0.09
    åĮ»åĬ¡äººåijĺ
    0.09
    åĩĿ
    0.08
    boxing
    0.08
    Activations Density 0.080%

    No Known Activations