© 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. 56097
    Prev
    Next
    INDEX
    Explanations

    ...unfortunateThis neuron detects conjunctions or follow-up phrases after ellipses

    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
    çĴľ
    -0.11
    天ä¸ĭ
    -0.10
     Cant
    -0.09
    ãģĮãĤī
    -0.09
    éĿŀæ´²
    -0.09
    åIJĮæł·æĺ¯
    -0.09
    .sn
    -0.08
    æĬ¤èº«ç¬¦
    -0.08
    æĪĸå¤ļ
    -0.08
    xic
    -0.08
    POSITIVE LOGITS
    å¾®å¦Ļ
    0.09
    Tit
    0.08
     retract
    0.08
    缯
    0.08
    ãĥĮ
    0.08
    zon
    0.08
    ye
    0.08
    رس
    0.08
    ISS
    0.08
     Tit
    0.08
    Activations Density 0.037%

    No Known Activations