© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 37598
    Prev
    Next
    INDEX
    Explanations

    encoding problems and symbols

    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
    綦
    -0.11
    åįij
    -0.10
    oug
    -0.10
    麻çĹ¹
    -0.10
     hypoc
    -0.10
    æ¯ħ
    -0.09
    æħĪ
    -0.09
    éļıæŶéļıåľ°
    -0.09
    çĽ¼
    -0.09
     Gund
    -0.09
    POSITIVE LOGITS
    ä¹±
    0.14
    äºĤ
    0.12
    ä¸ĢåĽ¢
    0.11
     unint
    0.11
     gib
    0.10
    éļıæľº
    0.10
    çľĭä¸įæĩĤ
    0.09
     gar
    0.09
    å±Ī
    0.09
     nonsense
    0.09
    Activations Density 0.165%

    No Known Activations