© 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. 37666
    Prev
    Next
    INDEX
    Explanations

    still, special, also, not, just, want, must

    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
    对äºİæĪij们
    -0.10
    ä¸įçĪ±
    -0.10
    ä¸įåıĺ
    -0.09
    æĺ¯æ²¡æľī
    -0.09
     dap
    -0.08
    AAP
    -0.08
    dale
    -0.08
    æĸ°å½¢åĬ¿
    -0.08
    nof
    -0.08
     Hers
    -0.08
    POSITIVE LOGITS
    erring
    0.09
    æ»ĭåij³
    0.09
    ran
    0.09
    esting
    0.09
    leans
    0.09
    orney
    0.08
    容æĺĵ
    0.08
    ä¹Łæĺ¯ä¸Ģç§į
    0.08
    sville
    0.08
    icked
    0.08
    Activations Density 0.196%

    No Known Activations