© 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. 1044
    Prev
    Next
    INDEX
    Explanations

    contrasting conjunctions or words introducing exceptions

    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
    åıªä¸įè¿ĩæĺ¯
    -0.11
    äºĪ
    -0.11
    ä¸įè¿ĩæĺ¯
    -0.11
    åĦĺ
    -0.10
    èĻ½è¯´
    -0.10
    æĶ¶èİ·
    -0.10
    è¿ĻäºĽéĹ®é¢ĺ
    -0.10
    çIJĨäºĭä¼ļ
    -0.09
    æĤ²ä¼¤
    -0.09
    èĻ½
    -0.09
    POSITIVE LOGITS
    å¦Ĥæŀľæĺ¯
    0.10
    pat
    0.09
    íĭĢ
    0.09
     Ø£Ùħا
    0.08
    nge
    0.08
    çİ©
    0.08
    اÙħا
    0.08
    å½ĵæĪij们
    0.08
    æĺŃ
    0.08
    å½ĵ
    0.08
    Activations Density 0.116%

    No Known Activations