© 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. 61495
    Prev
    Next
    INDEX
    Explanations

    remove, discard, return, wasted

    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
    upy
    -0.10
    ල
    -0.09
    æīĵæĸŃ
    -0.09
    åģ¶
    -0.09
    糯米
    -0.09
    æŀ¶
    -0.09
    Configurer
    -0.08
    ecute
    -0.08
    æĸĩåĮĸåĴĮ
    -0.08
    éķŃ
    -0.08
    POSITIVE LOGITS
    åı¦è¡Į
    0.16
    çķĻåŃĺ
    0.11
    çķĻä¸ĭæĿ¥
    0.11
    æĭ¿æĿ¥
    0.10
     discarded
    0.10
    åı¦ä¸Ģè¾¹
    0.10
    æĶ¶éĽĨ
    0.10
    æľīä»·åĢ¼çļĦ
    0.10
    åı¯çĶ¨äºİ
    0.10
    èĢķåľ°
    0.09
    Activations Density 0.118%

    No Known Activations