© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 37060
    Prev
    Next
    INDEX
    Explanations

    tech product and industry news

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron detects headings or section titles (tokens that appear in document or post titles/headers, often capitalized and near the start).

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31
    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
    è¿ĶåĽŀæIJľçĭIJ
    -0.10
     ...↵↵↵↵
    -0.09
    楼主
    -0.09
    ï¿
    -0.09
     PMID
    -0.09
    [email
    -0.08
    åįĹæµ·
    -0.08
     âĢ¦↵↵
    -0.08
    \\/
    -0.08
    %C
    -0.07
    POSITIVE LOGITS
    åĶ¿
    0.12
    ä¸ĩåįļ
    0.10
    QE
    0.09
    itest
    0.09
     POLITICO
    0.08
    YT
    0.08
    ENTA
    0.08
    ±Ð¾ÑĤ
    0.08
    æł¼å°ĩ
    0.08
    Ä°R
    0.08
    Activations Density 98.791%

    No Known Activations