© 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. 30376
    Prev
    Next
    INDEX
    Explanations

    sharp increase or decrease

    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
    è¿ĩçļĦ
    -0.11
    çĸĻ
    -0.09
    æ¯Ľå·¾
    -0.09
    an
    -0.08
    anse
    -0.08
    æĭħå½ĵ
    -0.08
     Sinn
    -0.08
    pecial
    -0.08
    of
    -0.08
    nt
    -0.07
    POSITIVE LOGITS
    arent
    0.11
    ستاÙĨ
    0.10
    ness
    0.10
    azon
    0.10
    alse
    0.10
    boards
    0.09
    èįª
    0.09
    OLL
    0.09
    lander
    0.09
    wares
    0.09
    Activations Density 0.005%

    No Known Activations