© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 18-TRANSCODER-HP
    7. 90216
    Prev
    Next
    INDEX
    Explanations

    food Method Num: 1, because all other methods failed

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_18.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.18.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    为空
    -0.30
    /catalog
    -0.29
    个å¤ļ
    -0.29
    utral
    -0.27
    locate
    -0.26
    Reviewed
    -0.25
    åªĴ
    -0.25
    缦
    -0.25
    underline
    -0.25
    åľ¨å®¶
    -0.24
    POSITIVE LOGITS
     Cons
    0.30
    jas
    0.28
    jos
    0.28
     cons
    0.27
     asymmetric
    0.27
     burning
    0.26
    cons
    0.26
    .wh
    0.25
    abar
    0.25
    ä½¥
    0.24
    Activations Density 8.440%

    No Known Activations