© 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. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 27726
    Prev
    Next
    INDEX
    Explanations

    "fine", "once", "computational knowledge engine", "compulsive decision", "Classic", "&&x", "Grass Roots", "Colligative property", "follow-only", "stimulatory budget"

    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
    ãģĹãģ¾ãģĨ
    -0.11
    lah
    -0.11
    à¹Į
    -0.09
    æĺ¯ä¸Ģ件
    -0.09
    åIJĮä¸Ģ个
    -0.09
    اء
    -0.09
    ctp
    -0.08
    aise
    -0.08
    æ½®
    -0.08
    Subsystem
    -0.08
    POSITIVE LOGITS
    ä¸įçķĻ
    0.11
    ä¿Ŀä½ı
    0.09
    ffer
    0.09
    ffective
    0.09
    ffect
    0.09
    åħ¬è®¤çļĦ
    0.09
    str
    0.09
    erator
    0.09
    æĸ°è¥¿
    0.09
    ìĹĶ
    0.09
    Activations Density 0.038%

    No Known Activations