© 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. 12182
    Prev
    Next
    INDEX
    Explanations

    `in` and `ir` prefixes with negative suffixes

    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
    สรร
    -0.10
    ero
    -0.09
    ropa
    -0.09
    ï¸
    -0.09
    isas
    -0.09
    åħ¶ä¸Ńä¸Ģ个
    -0.09
    åĴª
    -0.09
    ropy
    -0.08
    isable
    -0.08
    зо
    -0.08
    POSITIVE LOGITS
    å°ıäºĭ
    0.09
    /out
    0.09
    ä¸Ģä¼ļ
    0.09
    ä¸Ģé¡¿
    0.09
    åĩºä¼Ĺ
    0.09
    æĴ¼
    0.09
    ç³Ĭæ¶Ĥ
    0.09
    ç¼Ĭ
    0.09
    ç¡®åĪĩ
    0.08
    çĩĥ
    0.08
    Activations Density 0.055%

    No Known Activations