© 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. 64372
    Prev
    Next
    INDEX
    Explanations

    Technical/publication texts

    np_max-act · gemini-2.0-flash
    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
    éķ°
    -0.28
    º«
    -0.28
    umper
    -0.27
    åĩ¼
    -0.27
    iston
    -0.27
    PG
    -0.26
    身
    -0.26
    ç¢ĩ
    -0.26
    redi
    -0.26
     Engel
    -0.25
    POSITIVE LOGITS
    hashed
    0.27
    plain
    0.26
     recurs
    0.25
     ÐĴлад
    0.25
    rud
    0.25
    è¾¾
    0.24
    丢
    0.24
    lid
    0.24
     CWE
    0.24
    zÅij
    0.23
    Activations Density 50.377%

    No Known Activations