© 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. 33-TRANSCODER-HP
    7. 52745
    Prev
    Next
    INDEX
    Explanations

    Technical or foreign text

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    owl
    -0.29
    nesc
    -0.27
    ovo
    -0.26
    GNU
    -0.26
    ãĥį
    -0.25
    ICA
    -0.24
    ad
    -0.24
    _via
    -0.24
     verbally
    -0.24
    gnu
    -0.24
    POSITIVE LOGITS
    æĽ
    0.29
    steen
    0.28
    çĪĨ
    0.26
    å®ģæĦ¿
    0.26
    ä¸ĵ线
    0.26
     specialist
    0.26
    ели
    0.25
    鼷
    0.25
     Pod
    0.25
     cos
    0.24
    Activations Density 0.037%

    No Known Activations