© 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. 31577
    Prev
    Next
    INDEX
    Explanations

    patterns of behavior

    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
    UCH
    -0.11
    æĹłäºº
    -0.09
    人人
    -0.09
    linger
    -0.08
    å¤ĸåĩº
    -0.08
    .iso
    -0.08
    ä¸į失
    -0.08
    uch
    -0.08
    uart
    -0.08
     dart
    -0.08
    POSITIVE LOGITS
     pattern
    0.26
    pattern
    0.25
     patterns
    0.24
    è§Ħå¾ĭ
    0.21
     Patterns
    0.21
    ãĥijãĤ¿ãĥ¼ãĥ³
    0.21
    -pattern
    0.21
    patterns
    0.21
    è¶ĭåĬ¿
    0.20
    (pattern
    0.20
    Activations Density 0.063%

    No Known Activations