© 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. 53526
    Prev
    Next
    INDEX
    Explanations

    peeling off

    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
    sWith
    -0.10
    æŀģæĺĵ
    -0.09
     пÑĢид
    -0.09
     تعاÙĦÙī
    -0.09
    eur
    -0.09
    ted
    -0.09
    del
    -0.09
    des
    -0.08
    åĭĥåĭĥ
    -0.08
    gest
    -0.08
    POSITIVE LOGITS
    ander
    0.11
    è¶Ĭæĺ¯
    0.10
    ä½ķåĨµ
    0.10
    stakes
    0.10
    ings
    0.09
    etter
    0.09
    tures
    0.09
    鹫
    0.09
    ocks
    0.09
    inator
    0.09
    Activations Density 0.008%

    No Known Activations