© 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. 50081
    Prev
    Next
    INDEX
    Explanations

    "actual" followed by something concrete

    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
    çļĦä¸Ģ款
    -0.10
    æĽ´å¤ļä¿¡æģ¯
    -0.10
    latex
    -0.09
    篷
    -0.09
    ä¸įæ¸ħ
    -0.09
    eners
    -0.09
    pone
    -0.09
    quent
    -0.09
    bie
    -0.09
    lessness
    -0.09
    POSITIVE LOGITS
    ity
    0.16
    æĦıä¹īä¸ĬçļĦ
    0.13
    ités
    0.13
    mente
    0.11
    itus
    0.11
    ization
    0.10
    iable
    0.09
    ites
    0.09
     Downing
    0.09
    urd
    0.09
    Activations Density 0.032%

    No Known Activations