© 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. 38597
    Prev
    Next
    INDEX
    Explanations

    benign, harmless, or innocent

    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
    ãģĦãĤĭ
    -0.10
    æŀ¯
    -0.10
    اسÙĬ
    -0.09
    æħķ
    -0.09
    _AXIS
    -0.09
    ãĥģãĥ¥
    -0.09
    ØŃس
    -0.08
    .TH
    -0.08
    пÑĢÑı
    -0.08
    éªĦåĤ²
    -0.08
    POSITIVE LOGITS
     innoc
    0.14
     harmless
    0.13
     Harm
    0.11
     benign
    0.11
    éļĨ
    0.09
     innocent
    0.09
    alytics
    0.09
    ä¸įèµ·
    0.09
     Innoc
    0.09
    åĴĮæĶ¯æĮģ
    0.09
    Activations Density 0.043%

    No Known Activations