© 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. 20-TRANSCODER-HP
    7. 130315
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     originated
    -0.30
    ographies
    -0.28
    èĥĮæĻ¯
    -0.27
     originate
    -0.27
    hya
    -0.26
    娱ä¹IJåľº
    -0.26
    orpor
    -0.25
    ä¸Ģå¼ł
    -0.24
    åIJ¸å¼ķåĬĽ
    -0.24
    tems
    -0.24
    POSITIVE LOGITS
     Calories
    0.27
     mine
    0.25
     Lent
    0.25
     обÑıзан
    0.24
     cá»ij
    0.24
    isme
    0.24
     Sid
    0.24
    ense
    0.24
    å®ļåIJij
    0.23
    åĪĨæŀIJå¸Ī
    0.23
    Activations Density 8.921%

    No Known Activations

    This feature has no known activations.