© 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. 18050
    Prev
    Next
    INDEX
    Explanations

    strengths and weaknesses

    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
    æĮ¡
    -0.08
    ãģİ
    -0.08
    å¹²èĦĨ
    -0.08
    ç§ģèĩª
    -0.08
    ä»İä¸Ń
    -0.08
     Sculpt
    -0.08
    åŃĹæł·
    -0.08
    å¾Ĺ罪
    -0.07
    ç»Łè®¡æķ°æį®
    -0.07
    Ñħа
    -0.07
    POSITIVE LOGITS
     strengths
    0.17
    ä¼ĺçĤ¹
    0.16
    缺çĤ¹
    0.15
    å¼±çĤ¹
    0.14
    çļĦä¼ĺçĤ¹
    0.12
     weaknesses
    0.12
    çijķ
    0.12
    èĸĦå¼±
    0.11
    åĬ£åĬ¿
    0.11
    ä¼ĺåĬ¿
    0.10
    Activations Density 0.029%

    No Known Activations