© 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. 23538
    Prev
    Next
    INDEX
    Explanations

    attack or offensive noun

    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
    ually
    -0.11
    .getAs
    -0.09
    Ton
    -0.09
    irate
    -0.09
    getAs
    -0.08
    shal
    -0.08
    峦
    -0.08
    arResult
    -0.08
    è¾ĥé«ĺçļĦ
    -0.08
    WITHOUT
    -0.08
    POSITIVE LOGITS
    aine
    0.09
    APS
    0.09
    ä¸įéĹ´æĸŃ
    0.08
    ä¸įæĥľ
    0.08
    åħĪ导
    0.08
    kees
    0.08
    è¯ķæİ¢
    0.08
    ader
    0.08
     hitter
    0.08
    è§Ĩè§ī
    0.08
    Activations Density 0.027%

    No Known Activations