© 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. 35-TRANSCODER-HP
    7. 38824
    Prev
    Next
    INDEX
    Explanations

    "rel" or "less"

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_35.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.35.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     dipped
    -0.28
    ãĥ¥
    -0.27
    alah
    -0.26
    rams
    -0.25
    _SR
    -0.25
    drops
    -0.25
    æĬĵä½ı
    -0.25
    ollo
    -0.24
    bulk
    -0.24
    ä¸įçĪ±
    -0.24
    POSITIVE LOGITS
    lesia
    0.37
    legs
    0.36
    less
    0.36
    leg
    0.36
    åĪĹ
    0.36
     relate
    0.36
     relates
    0.36
     relevant
    0.34
    -less
    0.33
     relating
    0.33
    Activations Density 92.564%

    No Known Activations