© 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. 49742
    Prev
    Next
    INDEX
    Explanations

    common initial trigrams/digrams

    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
    ovo
    -0.09
    odiac
    -0.09
    è·
    -0.09
     бÑĥдÑĥ
    -0.09
    beat
    -0.08
    éĨĭ
    -0.08
    ãģ¬
    -0.08
    __/
    -0.08
    _attrib
    -0.08
    ingly
    -0.08
    POSITIVE LOGITS
    ove
    0.09
    ç¼ĺ
    0.08
    nage
    0.08
    r
    0.08
    dex
    0.08
    iding
    0.08
    æľĹ
    0.08
    iset
    0.08
    äºĽ
    0.08
    ä¸Ģ个å°ı
    0.07
    Activations Density 0.187%

    No Known Activations