© 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. 4493
    Prev
    Next
    INDEX
    Explanations

    adverbs of certainty or manner

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

    the neuron detects past-tense narrative language — verbs and markers used when recounting personal experiences.

    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
     kì
    -0.10
    Alright
    -0.10
     FUCK
    -0.09
     fucks
    -0.08
    è¿Ļ群
    -0.08
     terrifying
    -0.08
    群
    -0.08
    é¹Ħ
    -0.08
    ä¸Ģçŀ¬éĹ´
    -0.08
    umblr
    -0.08
    POSITIVE LOGITS
     cog
    0.09
    æĪijä¸įçŁ¥éģĵ
    0.09
    OLID
    0.09
    æĪij个人
    0.09
    parms
    0.09
     пенÑģи
    0.08
    èĶļ
    0.08
    _parm
    0.08
     sez
    0.08
    个人è§ĤçĤ¹
    0.08
    Activations Density 0.846%

    No Known Activations