© 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. 11784
    Prev
    Next
    INDEX
    Explanations

    level shifting or level of

    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
    uddy
    -0.10
    çķĮçļĦ
    -0.09
    ´
    -0.09
    lation
    -0.08
    ëĿ¼ëıĦ
    -0.08
    Ñĩик
    -0.08
    çľ¼éĩĮ
    -0.08
     foul
    -0.08
    ļ
    -0.08
    æĪIJç«ĭ
    -0.08
    POSITIVE LOGITS
    (Level
    0.13
    led
    0.13
    hood
    0.12
    .SEVERE
    0.11
    headed
    0.10
    -headed
    0.10
     serialVersionUID
    0.10
    à¸ĩาà¸Ļ
    0.10
    =logging
    0.10
    .setLevel
    0.09
    Activations Density 0.090%

    No Known Activations