© 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. 32064
    Prev
    Next
    INDEX
    Explanations

    black humour, wit, comedy

    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
    fel
    -0.10
    éĩįéĩį
    -0.09
    oku
    -0.08
    累累
    -0.08
    ç»ī
    -0.08
    åģİ
    -0.08
    åľ°ä½į
    -0.08
    å¯IJ
    -0.08
    ums
    -0.08
    å¤ļå¤ļ
    -0.08
    POSITIVE LOGITS
    å¹½é»ĺ
    0.13
     humour
    0.13
    éĤ£ä»½
    0.12
     humor
    0.12
    ä¸Ģ份
    0.11
     sense
    0.10
    使åij½æĦŁ
    0.10
    åįģ足
    0.10
    æĥ³è±¡åĬĽ
    0.10
    è¿Ļ份
    0.10
    Activations Density 0.047%

    No Known Activations