© 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. 10521
    Prev
    Next
    INDEX
    Explanations

    letter or letters

    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
    éķ¿éĢĶ
    -0.09
    åħ´ä¸ļ
    -0.08
    odash
    -0.08
    éĵĥ
    -0.08
    æ´ģ
    -0.08
    POCH
    -0.08
    ridged
    -0.08
    rio
    -0.08
    éĵºè®¾
    -0.08
     Heg
    -0.08
    POSITIVE LOGITS
    (letter
    0.15
     letters
    0.14
     letter
    0.14
    letter
    0.14
    -letter
    0.13
    åŃĹæ¯į
    0.12
     Letters
    0.12
    _letters
    0.12
    Letter
    0.11
    _letter
    0.11
    Activations Density 0.075%

    No Known Activations