© 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. 61812
    Prev
    Next
    INDEX
    Explanations

    negative characteristics

    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
    ย
    -0.13
    ized
    -0.13
    ãģĹãģı
    -0.10
    çĦ¶æĺ¯
    -0.09
    ised
    -0.09
    IZED
    -0.09
    ize
    -0.09
    æIJ
    -0.09
    izer
    -0.09
    izing
    -0.09
    POSITIVE LOGITS
    inator
    0.12
    MENT
    0.11
    ingly
    0.10
    ement
    0.10
    ments
    0.10
    ers
    0.10
    von
    0.10
    antly
    0.10
    ernet
    0.09
    sWith
    0.09
    Activations Density 0.075%

    No Known Activations