© 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. 27900
    Prev
    Next
    INDEX
    Explanations

    home, Copyright, React, old

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

    The neuron detects vivid, poetic/lyrical language—emotionally charged descriptive imagery and rhetorical, metaphorical phrasing.

    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
     Äijiá»ĥn
    -0.09
    åı¯è°ĵæĺ¯
    -0.09
    reon
    -0.09
    磼
    -0.09
    èĽĨ
    -0.09
    æľīä¸įå°ij
    -0.09
    å¢ŀ
    -0.09
    aida
    -0.09
    é²ľèĬ±
    -0.08
    .pad
    -0.08
    POSITIVE LOGITS
    irut
    0.09
    omat
    0.09
     пÑĢож
    0.08
    /Framework
    0.08
     res
    0.08
    ãģ¯ãģĺãĤģ
    0.07
    ieres
    0.07
     fucked
    0.07
     myth
    0.07
     accr
    0.07
    Activations Density 0.216%

    No Known Activations