© 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. 4735
    Prev
    Next
    INDEX
    Explanations

    narrow or narrower

    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
    nice
    -0.09
    满æĢĢ
    -0.08
    Ñıвление
    -0.08
    ugu
    -0.08
    entic
    -0.08
    çµ®
    -0.08
    ода
    -0.08
     nice
    -0.08
    ething
    -0.08
    çĽijäºĭä¼ļ
    -0.07
    POSITIVE LOGITS
    çĭŃ
    0.12
    çªĦ
    0.12
    åľ°å¸¦
    0.12
    -minded
    0.12
     minded
    0.11
    -mouth
    0.11
    éĻ¬
    0.10
    éļĺ
    0.10
    çĭŃçªĦ
    0.10
    ling
    0.10
    Activations Density 0.013%

    No Known Activations