© 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. 61161
    Prev
    Next
    INDEX
    Explanations

    "dis" followed by words starting with "am", "claim", "ability", "emb"

    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
    ä¸Ģå¼Ģå§ĭå°±
    -0.11
    èŀ¨
    -0.11
    ä¹Łéļıä¹ĭ
    -0.10
    sWith
    -0.10
    æĭĬ
    -0.09
    ç°Į
    -0.09
    æ·¼
    -0.09
    itud
    -0.09
    åĺ´å·´
    -0.09
    ä¸Ģ缴
    -0.09
    POSITIVE LOGITS
    bursement
    0.10
    /dis
    0.10
    ble
    0.09
    rances
    0.09
    .dis
    0.09
    cription
    0.09
    ertation
    0.09
    asaki
    0.09
    (dis
    0.09
    r
    0.08
    Activations Density 0.035%

    No Known Activations