© 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. 56286
    Prev
    Next
    INDEX
    Explanations

    words ending in an followed by specific suffixes

    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
    ت
    -0.12
    vo
    -0.12
    ãģĵãģ¨ãģĮãģ§ãģįãģ¾ãģĻ
    -0.11
    elin
    -0.10
    gee
    -0.09
    ष
    -0.09
    ger
    -0.09
    el
    -0.09
    Ñģка
    -0.09
    cek
    -0.09
    POSITIVE LOGITS
    terior
    0.14
    omaly
    0.12
    rans
    0.12
    ifold
    0.11
    berra
    0.11
    tern
    0.11
    alysis
    0.11
    enko
    0.11
    ufact
    0.11
    ãģģ
    0.10
    Activations Density 0.154%

    No Known Activations