© 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. 22080
    Prev
    Next
    INDEX
    Explanations

    always or never

    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
    éŨå¤ĸ
    -0.08
     boot
    -0.08
    rut
    -0.08
    ëĭ¤ë©´
    -0.08
    .cls
    -0.08
     âĶ
    -0.08
    ä½Ĩå¦Ĥæŀľ
    -0.08
    /AP
    -0.08
    主管
    -0.08
     Mor
    -0.07
    POSITIVE LOGITS
    ä»»ä½ķæŶåĢĻ
    0.19
    å§ĭç»Ī
    0.19
     always
    0.18
     вÑģегда
    0.17
    æĢ»æĺ¯
    0.17
    always
    0.17
    常ãģ«
    0.15
    æŶåĪ»
    0.15
    å§ĭç»Īä¿ĿæĮģ
    0.15
     ALWAYS
    0.15
    Activations Density 0.019%

    No Known Activations