© 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. 26720
    Prev
    Next
    INDEX
    Explanations

    hook and loop

    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
    -prefix
    -0.10
    èģĶ
    -0.10
    åķľ
    -0.09
    æ¦ĤåĨµ
    -0.08
    eming
    -0.08
     wor
    -0.08
    éĶ¦ç»£
    -0.08
    _te
    -0.08
    .hex
    -0.08
     Restoration
    -0.08
    POSITIVE LOGITS
    anz
    0.10
    æĶ¶ç´§
    0.10
     makeup
    0.09
     недо
    0.09
     khóa
    0.09
    ihad
    0.09
    owers
    0.08
    客è§Ĥ
    0.08
    å¾ģæĶ¶
    0.08
    ·¨
    0.08
    Activations Density 0.110%

    No Known Activations