© 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. Qwen3-4B
    5. 32-TRANSCODER-HP
    7. 43642
    Prev
    Next
    INDEX
    Explanations

    technical documentation/configurations

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æĺ¯æ²¡æľī
    -0.27
    æĺ¯æ¯Ķè¾ĥ
    -0.26
    ä»İåīį
    -0.26
    phia
    -0.25
    å°±æľīäºĨ
    -0.24
    -outs
    -0.24
    æĦı大
    -0.24
    ahas
    -0.23
    ç«ŀ
    -0.23
    åIJ«
    -0.23
    POSITIVE LOGITS
    åİ»
    0.52
    æĿ¥è¿Ľè¡Į
    0.33
     Deal
    0.33
    æĿ¥
    0.31
     deal
    0.31
     effectively
    0.31
    Deal
    0.30
     correctly
    0.29
    è¦ģåİ»
    0.29
    å¦Ĥä½ķåİ»
    0.29
    Activations Density 0.000%

    No Known Activations