© 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. 113039
    Prev
    Next
    INDEX
    Explanations

    :

    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
    uant
    -0.26
    éĿ¢å¯¹
    -0.25
    èıĬèĬ±
    -0.24
    æĭħ
    -0.24
    è¦ģæľī
    -0.23
    ä¸Ĭä¼ł
    -0.23
    Łèĥ½
    -0.23
    éŃĶæľ¯
    -0.23
    rail
    -0.23
    onic
    -0.23
    POSITIVE LOGITS
     reproduce
    0.24
    Child
    0.24
     accomp
    0.23
    èĩªæĿ¥
    0.23
    Fi
    0.23
    -NLS
    0.23
    .sd
    0.22
    媳
    0.22
     VOL
    0.22
    elyn
    0.22
    Activations Density 0.001%

    No Known Activations