© 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. 22951
    Prev
    Next
    INDEX
    Explanations

    had to / must + verb

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    åĿĿ
    -0.10
    ä¸įæĢķ
    -0.09
    unched
    -0.08
    èĵ¬åĭĥ
    -0.08
    _TIM
    -0.08
    drawing
    -0.08
    ä¸ļ
    -0.08
    æĭīä¸ģ
    -0.07
    æŃ£å½ĵ
    -0.07
     Wor
    -0.07
    POSITIVE LOGITS
    åľ¨ä¸Ģ个
    0.10
    ARN
    0.10
    æ²½
    0.09
    çĦ¶æĺ¯
    0.09
     asn
    0.09
    unist
    0.09
    幺
    0.09
    é¹Ĺ
    0.09
    å¹²é¢Ħ
    0.08
    -neck
    0.08
    Activations Density 0.027%

    No Known Activations