© 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. 26697
    Prev
    Next
    INDEX
    Explanations

    many/few features/objects/measures

    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
    Cause
    -0.08
    çĩĥ
    -0.08
    éĹª
    -0.08
    æĿł
    -0.08
    è·³
    -0.08
    好好
    -0.08
    渣
    -0.08
    éĴ±
    -0.08
    æīĢ以æĪij
    -0.08
    éħį
    -0.08
    POSITIVE LOGITS
    åħ¶ä»ĸçļĦ
    0.11
    é¢Ŀå¤ĸ
    0.11
    icc
    0.10
     optional
    0.10
    éĢī项
    0.10
    è¾ħåĬ©
    0.10
    noon
    0.10
    обÑĢаз
    0.09
    leet
    0.09
    åIJĦ项
    0.09
    Activations Density 0.245%

    No Known Activations