© 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. 33006
    Prev
    Next
    INDEX
    Explanations

    offended, sensitive, or hurt

    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
    çŃĭ
    -0.10
    æŃ¢
    -0.10
    æĺİæĺİ
    -0.09
    çªģåĩº
    -0.08
    æĿĤ
    -0.08
    èµŀåı¹
    -0.08
    IDO
    -0.08
    âĨĵ
    -0.08
    -t
    -0.08
    vor
    -0.08
    POSITIVE LOGITS
    æķıæĦŁ
    0.17
    åĤ·
    0.15
     sensit
    0.15
    伤害
    0.15
    伤
    0.14
     hurt
    0.14
    伤å¿ĥ
    0.13
    çģµæķı
    0.13
    åıĹ伤
    0.13
     sensitivity
    0.12
    Activations Density 0.049%

    No Known Activations