© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 21505
    Prev
    Next
    INDEX
    Explanations

    loss or waste

    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
    ä¸įä¸Ģå®ļ
    -0.10
    æĥļ
    -0.09
    ä¸įå®ļ
    -0.09
    çħ½
    -0.09
    ä¸ī人
    -0.09
    æģĭ人
    -0.08
    ä¸¤åĽ½
    -0.08
    æĽ´å®¹æĺĵ
    -0.08
    ä»ĭåħ¥
    -0.08
    åıįèħIJåĢ¡å»ī
    -0.08
    POSITIVE LOGITS
    浪费
    0.18
    å®Ŀè´µçļĦ
    0.16
     wasted
    0.15
    çĻ½çĻ½
    0.14
    å®Ŀè´µ
    0.13
    æ½ľåĬĽ
    0.13
    åı¯æĥľ
    0.12
    æ½ľèĥ½
    0.12
    æľºä¼ļ
    0.11
    çıįè´µ
    0.11
    Activations Density 0.104%

    No Known Activations