© 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. 10089
    Prev
    Next
    INDEX
    Explanations

    "others" followed by contrasting actions

    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»ĸä¹Ł
    -0.10
    æĬ»
    -0.10
    è¿Ļ家ä¼Ļ
    -0.10
    æĻļéĹ´
    -0.09
    çĻ½å¤©
    -0.09
    lest
    -0.09
    ä¸įåĨįæĺ¯
    -0.09
    ä¹Łæĺ¯ä¸Ģ个
    -0.08
    åĿ¨
    -0.08
    诱人
    -0.08
    POSITIVE LOGITS
    çĶļèĩ³æĺ¯
    0.13
    çĶļèĩ³
    0.12
     вообÑīе
    0.11
     пÑĢоÑģÑĤо
    0.10
    æĽ´éĩįè¦ģ
    0.10
    åĪĻæĺ¯
    0.09
    åĽłåľ°
    0.09
     Emb
    0.09
    emb
    0.09
    .Css
    0.09
    Activations Density 0.043%

    No Known Activations