© 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. 57808
    Prev
    Next
    INDEX
    Explanations

    punctuation followed by a word

    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
    æĬ±çĿĢ
    -0.11
    ãĤ»ãĥĥãĥĪ
    -0.10
    ÙĪÙĦÙĬ
    -0.09
    æľĢåħ³éĶ®
    -0.09
    (*)(
    -0.09
    çĤ¬
    -0.09
    ACHED
    -0.09
    è¶ĬæĿ¥è¶Ĭå°ij
    -0.08
    æ¹ĸæ³Ĭ
    -0.08
    è¶ĬæĿ¥è¶Ĭ
    -0.08
    POSITIVE LOGITS
    ä¸İåIJ¦
    0.10
     же
    0.09
    æ¿Ģæĥħ
    0.09
    tps
    0.08
    ?
    0.08
     meanwhile
    0.08
     significa
    0.08
    ä¸įæĢķ
    0.08
    éĺħ读åħ¨æĸĩ
    0.08
    erton
    0.07
    Activations Density 0.138%

    No Known Activations