© 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. 62780
    Prev
    Next
    INDEX
    Explanations

    ) followed by "Field" or "Technical"

    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
    e
    -0.09
    o
    -0.09
    r
    -0.08
    born
    -0.08
    er
    -0.08
    ledge
    -0.08
    éĨĭ
    -0.08
    l
    -0.08
     Ùħج
    -0.08
    b
    -0.08
    POSITIVE LOGITS
    unts
    0.11
     gái
    0.09
    ä¸Ģåī¯
    0.09
     Ø¥ÙĦÙĬ
    0.09
    itional
    0.08
    vertising
    0.08
    unge
    0.08
    ìĹĶ
    0.08
    olis
    0.08
    sembly
    0.08
    Activations Density 0.139%

    No Known Activations