© 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. 62532
    Prev
    Next
    INDEX
    Explanations

    but/and followed by which/with/the

    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
     Äijè
    -0.09
    èį
    -0.09
    oned
    -0.08
    æĪij们ä¼ļ
    -0.08
    çŁ³æ¦´
    -0.08
    å¼ķãģ£
    -0.08
    è¿Ļæĺ¯ä¸Ģ个
    -0.08
     shack
    -0.08
    ä»ĸ们æĺ¯
    -0.08
    秣
    -0.08
    POSITIVE LOGITS
     whose
    0.16
    whose
    0.12
     ÙĪاÙĦتÙĬ
    0.12
    åıĪ被
    0.10
    è¿ĽèĢĮ
    0.10
    zek
    0.09
    ä½Ĩåį´
    0.09
     ÙĪاÙĦØ°ÙĬ
    0.09
    被称为
    0.09
     otherwise
    0.09
    Activations Density 0.261%

    No Known Activations