© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 16907
    Prev
    Next
    INDEX
    Explanations

    arrogant, own, fame, spoiled

    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
    æĵĺ
    -0.10
    å¿ĥçĸ¼
    -0.10
    æĭĸ延
    -0.09
    ä¸įåĩºæĿ¥
    -0.09
    æºIJæºIJ
    -0.09
    ä¸ĢéĶ®
    -0.09
    åħ³åĪĩ
    -0.09
     Fog
    -0.08
     Matte
    -0.08
    æĦıæĦ¿
    -0.08
    POSITIVE LOGITS
    éªĦ
    0.16
    éªĦåĤ²
    0.14
     ego
    0.13
    趾
    0.13
    åĤ²
    0.12
     arrogance
    0.12
     arrogant
    0.12
    çĤ«èĢĢ
    0.11
    è°¦
    0.11
    å¾ĹæĦı
    0.11
    Activations Density 0.046%

    No Known Activations