© 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. 17737
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    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
    htag
    -0.28
    BAT
    -0.27
     bat
    -0.26
     Poison
    -0.26
    æīĵèµ¢
    -0.25
     booth
    -0.25
     agree
    -0.25
    ismet
    -0.25
    ERVE
    -0.25
    _spinner
    -0.24
    POSITIVE LOGITS
    itudes
    0.29
    为客æĪ·
    0.29
     Cong
    0.27
    CONDS
    0.26
    Cons
    0.24
     applicationContext
    0.24
    lords
    0.24
    äºĮæĪĺ
    0.24
    è°Ľ
    0.23
     Conditioning
    0.23
    Activations Density 4.068%

    No Known Activations

    This feature has no known activations.