© 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. 30-TRANSCODER-HP
    7. 96936
    Prev
    Next
    INDEX
    Explanations

    Originally

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     consideration
    -0.35
    mutable
    -0.28
    uria
    -0.28
    IBC
    -0.27
     advice
    -0.27
     variation
    -0.26
    æĮĩ导æĦıè§ģ
    -0.26
    erde
    -0.26
    IRROR
    -0.25
     Variation
    -0.25
    POSITIVE LOGITS
     meant
    0.31
     formed
    0.31
     intended
    0.29
     conceived
    0.28
    Pro
    0.28
    设å®ļ
    0.27
     dreamed
    0.26
    åģĩ
    0.25
     developed
    0.25
    owitz
    0.25
    Activations Density 0.902%

    No Known Activations