© 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. 18897
    Prev
    Next
    INDEX
    Explanations

    "well" followed by punctuation or "my", "relates", "how", "2011"

    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
    æ´¾
    -0.09
    inos
    -0.09
    ä¸İåIJ¦
    -0.08
    èķ¨
    -0.08
    ripe
    -0.08
     muj
    -0.08
    IDS
    -0.08
    ä¸Ģæīĭ
    -0.08
    个åĪ«
    -0.08
    åĸīåĴĻ
    -0.08
    POSITIVE LOGITS
    让æĤ¨
    0.09
    å®ŀåľ¨æĺ¯å¤ª
    0.09
     Colbert
    0.09
    logg
    0.08
    åĩłä¹İæīĢæľī
    0.08
    æĢ»ä¹ĭ
    0.08
    æ¯ıå½ĵ
    0.08
    请æĤ¨
    0.08
    å®ŀåľ¨å¤ª
    0.08
    è¹Ī
    0.08
    Activations Density 0.031%

    No Known Activations