© 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. 31-TRANSCODER-HP
    7. 19707
    Prev
    Next
    INDEX
    Explanations

    BLE

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    æıĦ
    -0.32
    ,eg
    -0.27
     pud
    -0.27
    第ä¸ĢèĬĤ
    -0.26
    IRT
    -0.25
     StatusBar
    -0.25
    	sf
    -0.25
    è¿Ļæĺ¯æĪij
    -0.24
    SR
    -0.24
    /details
    -0.24
    POSITIVE LOGITS
    _traits
    0.27
    uyo
    0.27
    ÑĨов
    0.26
    corn
    0.25
    éķ°
    0.25
    á»ĥn
    0.25
     devant
    0.25
     oil
    0.25
    awe
    0.25
     Oil
    0.24
    Activations Density 0.008%

    No Known Activations