© 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. 52970
    Prev
    Next
    INDEX
    Explanations

    make something

    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
    hoot
    -0.10
    åģļ个
    -0.09
    /out
    -0.09
    åģļå¾Ĺ
    -0.09
    reed
    -0.09
    inas
    -0.09
    /on
    -0.09
    ighter
    -0.09
    öst
    -0.09
    rede
    -0.08
    POSITIVE LOGITS
     sure
    0.18
    leine
    0.16
     sense
    0.13
    äºĨä¸Ģåľº
    0.12
    -bel
    0.12
    overs
    0.11
    äºĨä¸Ģ
    0.11
    ⽤
    0.11
    RelativeTo
    0.11
    intosh
    0.10
    Activations Density 0.141%

    No Known Activations