© 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. 26902
    Prev
    Next
    INDEX
    Explanations

    apologize or apology

    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
    (tm
    -0.09
    icut
    -0.08
     rebel
    -0.08
    ç¨³åĽº
    -0.08
     пÑĢеимÑĥÑīеÑģÑĤв
    -0.08
     vat
    -0.08
    éĶģå®ļ
    -0.08
    éļ¾éģĵ
    -0.08
    hma
    -0.08
    yles
    -0.08
    POSITIVE LOGITS
    æĬ±æŃī
    0.23
    éģĵæŃī
    0.22
    æŃī
    0.21
     apology
    0.18
     apologies
    0.17
     sorry
    0.17
    sorry
    0.17
    对ä¸įèµ·
    0.17
    å¿ı
    0.16
     apologize
    0.16
    Activations Density 0.046%

    No Known Activations