© 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. 2377
    Prev
    Next
    INDEX
    Explanations

    complex phenomenon extends beyond

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

    words describing trends of growth, expansion, or increasing prevalence over time.

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

    The neuron detects words and tokens that indicate complex, systemic, or large‑scale/technical concepts (academic or scientific discourse vocabulary).

    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
    åħ¶å®ĥ
    -0.09
    åı¦å¤ĸä¸Ģ个
    -0.09
    2
    -0.08
    åħ¶æīĢ
    -0.08
    å¦ĤæѤ
    -0.08
    æĽ´å¥½
    -0.08
    3
    -0.08
    6
    -0.08
    1
    -0.08
    7
    -0.08
    POSITIVE LOGITS
    æİł
    0.09
    allery
    0.09
    ç§ijåĪĽæĿ¿
    0.09
    allas
    0.09
    é¦ĸè½®
    0.09
    é©Ń
    0.08
    å´½
    0.08
    ⽬
    0.08
    ESCO
    0.08
    _preferences
    0.08
    Activations Density 0.694%

    No Known Activations