© 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. 17945
    Prev
    Next
    INDEX
    Explanations

    the

    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
    注æĺİåĩºå¤Ħ
    -0.30
    为ä¹ĭ
    -0.26
    agger
    -0.25
    istogram
    -0.24
     taxable
    -0.24
     dbHelper
    -0.24
    herent
    -0.24
    atters
    -0.24
     even
    -0.23
     regular
    -0.23
    POSITIVE LOGITS
    è¿Ļä¸ī个
    0.29
    åĽ½éĻħåĮĸ
    0.28
    è¿Ļ个
    0.28
     blanc
    0.27
    大éĻĨ
    0.27
     grub
    0.26
    æŀ¶
    0.25
     Tower
    0.25
    空éĹ´
    0.25
     Statue
    0.25
    Activations Density 0.584%

    No Known Activations