© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 24257
    Prev
    Next
    INDEX
    Explanations

    enough to bringenough to saveenough to allowenough to preventenough to competeenough to makeenough to provokeenough to raiseenough to send

    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
    Negative Logits
    ä¸ĬäºĨ
    -0.09
    å·²ä¹ħ
    -0.09
    åĢľ
    -0.09
    以人为
    -0.09
    äºĨèµ·æĿ¥
    -0.09
    abouts
    -0.09
    說話
    -0.08
     own
    -0.08
     Wrest
    -0.08
    åĤij
    -0.08
    POSITIVE LOGITS
    带åĬ¨
    0.09
    æij
    0.09
    çľ¼çļ®
    0.08
    .tm
    0.08
    ERVE
    0.08
    èµĦ
    0.08
    yt
    0.08
    ä¹°åįķ
    0.08
    对éĿ¢
    0.08
    ÐĺÐĿ
    0.08
    Activations Density 0.025%

    No Known Activations