© 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. 32670
    Prev
    Next
    INDEX
    Explanations

    de facto, effectively, effect

    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
    æŃī
    -0.09
    åĭ¤
    -0.09
    æĦŁæģ©
    -0.09
    æľĢå¤ļ
    -0.08
    åĽŀæĬ¥
    -0.08
    LineStyle
    -0.08
    acco
    -0.08
    çīĩ
    -0.08
    åĸĥ
    -0.08
    强åĬĽ
    -0.08
    POSITIVE LOGITS
    æĦıä¹īä¸ĬçļĦ
    0.14
    æĦıä¹īä¸Ĭ
    0.13
    ãģ»ãģ¼
    0.11
    æĿ¥è®²
    0.11
    æĿ¥è¯´
    0.10
    icial
    0.10
     essentially
    0.10
    å®ŀè´¨
    0.10
    èĢĮè¨Ģ
    0.10
     purposes
    0.10
    Activations Density 0.036%

    No Known Activations