© 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. 22-TRANSCODER-HP
    7. 128885
    Prev
    Next
    INDEX
    Explanations

    Describing a process

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    Anywhere
    -0.27
    ç¾İ人
    -0.26
    expected
    -0.26
    é¢ĦæľŁ
    -0.25
    erman
    -0.24
    lesi
    -0.24
     Abr
    -0.24
    	expected
    -0.24
    anes
    -0.23
    ÙħÙĤاÙĦات
    -0.23
    POSITIVE LOGITS
     consists
    0.38
     involves
    0.31
     consist
    0.30
     focuses
    0.29
     follows
    0.26
     consisted
    0.26
     includes
    0.25
     recognizes
    0.25
     navigator
    0.25
    åĪĨ为
    0.25
    Activations Density 0.049%

    No Known Activations