© 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. 31-TRANSCODER-HP
    7. 19822
    Prev
    Next
    INDEX
    Explanations

    problems/difficulties

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    ITU
    -0.24
    vara
    -0.24
     BEEN
    -0.23
     PTS
    -0.23
    composite
    -0.23
    为主è¦ģ
    -0.23
    岬
    -0.23
    enstein
    -0.23
    tape
    -0.23
    è½½ä½ĵ
    -0.22
    POSITIVE LOGITS
     connected
    0.52
    connected
    0.44
     connecting
    0.44
    è¿ŀæİ¥
    0.43
     Connected
    0.39
     linked
    0.38
    Connected
    0.37
    -connected
    0.37
     Connecting
    0.35
    _connected
    0.35
    Activations Density 0.176%

    No Known Activations