© 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. 54820
    Prev
    Next
    INDEX
    Explanations

    defending or definitive contexts

    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
    umph
    -0.10
    fully
    -0.10
    иÑĢовка
    -0.09
    iffs
    -0.09
    ¦
    -0.09
    åĵĩ
    -0.09
    è·
    -0.09
    วาà¸ĩ
    -0.09
    åĮľ
    -0.09
    vine
    -0.08
    POSITIVE LOGITS
    iveness
    0.13
    ective
    0.12
    eated
    0.12
    ÑģÑı
    0.11
    ragment
    0.11
    iance
    0.11
    initely
    0.10
    -minded
    0.10
    ä¸įäºĨ
    0.10
    rag
    0.10
    Activations Density 0.087%

    No Known Activations