© 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. 63662
    Prev
    Next
    INDEX
    Explanations

    Tokenizing text

    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
    nal
    -0.29
    å£ģ
    -0.26
    åĿļæŀľ
    -0.26
    led
    -0.24
    æ¯ķç«Ł
    -0.24
    ê¹¥
    -0.24
    LED
    -0.23
    /controller
    -0.23
     emotion
    -0.23
    jour
    -0.23
    POSITIVE LOGITS
    .bc
    0.26
    è¢ĸ
    0.25
     voices
    0.24
    vo
    0.24
    -)
    0.24
     derives
    0.23
     succeeds
    0.23
    aggio
    0.23
     deriving
    0.23
     impe
    0.23
    Activations Density 5.907%

    No Known Activations