© 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. 26775
    Prev
    Next
    INDEX
    Explanations

    punctuation questioning surprise

    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
    venue
    -0.09
    uell
    -0.09
     Ñģвоими
    -0.08
    /or
    -0.08
    owa
    -0.08
    âľµ
    -0.08
    idia
    -0.08
    unar
    -0.08
    æ¢Ĺ
    -0.08
    âĿĿ
    -0.08
    POSITIVE LOGITS
    leness
    0.10
    ities
    0.10
    ocket
    0.10
    åijĺ
    0.10
    PLE
    0.09
    izr
    0.09
    çĦ¶æĺ¯
    0.09
    åĴĮåıijå±ķ
    0.09
    ilda
    0.09
    ỡ
    0.09
    Activations Density 0.016%

    No Known Activations