© 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. 61910
    Prev
    Next
    INDEX
    Explanations

    prefix

    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
    spot
    -0.11
    engkap
    -0.09
    éĨĮ
    -0.09
    linky
    -0.09
    ÙĪاس
    -0.09
    -DD
    -0.09
     Pitch
    -0.09
    ran
    -0.09
    绽æĶ¾
    -0.08
    anh
    -0.08
    POSITIVE LOGITS
    ç¼Ģ
    0.11
    prefix
    0.09
    gio
    0.09
    (prefix
    0.09
    éĻĢ
    0.08
    举
    0.08
     prefix
    0.08
    帽
    0.08
    PathComponent
    0.08
     Mor
    0.08
    Activations Density 0.137%

    No Known Activations