© 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. 58666
    Prev
    Next
    INDEX
    Explanations

    spot or spots

    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
    hart
    -0.13
    æĸ¯é¡¿
    -0.11
    çģ¶
    -0.10
    sport
    -0.10
    çĵ£
    -0.09
    éĶ£
    -0.09
    oplast
    -0.09
    ance
    -0.09
    误导
    -0.09
    hq
    -0.08
    POSITIVE LOGITS
    ter
    0.13
    ween
    0.11
    åĦ¿
    0.10
    nett
    0.10
    andard
    0.10
    же
    0.09
    ting
    0.09
    åħĴ
    0.09
    еÑĩ
    0.09
    髯
    0.09
    Activations Density 0.014%

    No Known Activations