© 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. 32771
    Prev
    Next
    INDEX
    Explanations

    find string or replace

    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
    è£Ļ
    -0.10
    spy
    -0.09
    çĽ
    -0.08
    æĭ¢
    -0.08
    åłĨ积
    -0.08
    å®ĺæĸ¹å¾®åįļ
    -0.08
    =default
    -0.08
    æĶĴ
    -0.08
    éĺ²çĽĹ
    -0.08
     Manhattan
    -0.08
    POSITIVE LOGITS
    æľī害
    0.11
    needle
    0.11
    æĥ³æī¾
    0.10
    æī¾åĩº
    0.10
    å®īæĬļ
    0.10
    纫
    0.10
     needle
    0.10
    éĴĪ
    0.10
    æŁ¥æī¾
    0.09
    å¹»æĥ³
    0.09
    Activations Density 0.189%

    No Known Activations