© 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. 33844
    Prev
    Next
    INDEX
    Explanations

    French culture and cinema

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

    informational text passages that introduce and explain topics systematically.

    oai_token-act-pair · claude-4-5-haikuTriggered by @yooniel31

    tokens that carry topical or content meaning (important nouns, verbs, and other informative words).

    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
    为æĪij们
    -0.09
    iola
    -0.09
    å½ĵæĪij们
    -0.09
    ä¸ĢåĪĨ
    -0.09
    åĴ±
    -0.09
     folks
    -0.08
    为大家
    -0.08
    ãģ¯ãĤĤãģ¡ãĤįãĤĵ
    -0.08
    ä»Ĭå¹´
    -0.08
    å¤ļå¤ļ
    -0.08
    POSITIVE LOGITS
    [c
    0.13
    .Template
    0.11
    éĥ¨åĪĨåľ°åĮº
    0.10
     Template
    0.10
    èĩ³ä¸Ĭ
    0.09
    éĥ¨ä»½
    0.09
    ä¸ŃåĽ½å¤§éĻĨ
    0.09
    Template
    0.09
    ä½ĨçĶ±äºİ
    0.09
    ,[
    0.09
    Activations Density 5.952%

    No Known Activations