© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 51411
    Prev
    Next
    INDEX
    Explanations

    would, could, did, was, were

    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
    å¤Ķ
    -0.12
    audi
    -0.09
    ales
    -0.09
    Öģ
    -0.09
    åIJī
    -0.09
    èŃĺ
    -0.09
    è¯Ĩ
    -0.08
    kok
    -0.08
    åIJİæľŁ
    -0.08
    绳
    -0.08
    POSITIVE LOGITS
     would
    0.15
     Would
    0.13
    would
    0.13
     wouldn
    0.13
    就好äºĨ
    0.13
    Would
    0.13
     could
    0.13
    could
    0.12
     zou
    0.12
     fos
    0.11
    Activations Density 0.092%

    No Known Activations