© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 30-TRANSCODER-HP
    7. 96333
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    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
    Not in Any Lists

    No Comments

    Negative Logits
    voices
    -0.28
    elize
    -0.27
     SEEK
    -0.26
    :convert
    -0.26
    phia
    -0.26
     Liberties
    -0.26
    cgi
    -0.25
    lesia
    -0.25
     lesbian
    -0.25
    åĪ¶èį¯
    -0.25
    POSITIVE LOGITS
    inden
    0.32
    term
    0.30
    E
    0.28
     R
    0.28
     E
    0.27
     cod
    0.26
    ucket
    0.26
     Cod
    0.25
    ä¸įçŁ¥
    0.25
    inf
    0.25
    Activations Density 0.001%

    No Known Activations

    This feature has no known activations.