© 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. 31-TRANSCODER-HP
    7. 19027
    Prev
    Next
    INDEX
    Explanations

    can

    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
    Not in Any Lists

    No Comments

    Negative Logits
    éľĢè¦ģ
    -0.29
    lrt
    -0.28
    	parameters
    -0.27
    dio
    -0.26
     incons
    -0.26
     needs
    -0.26
     necesita
    -0.25
    丫
    -0.25
     paran
    -0.24
    cba
    -0.24
    POSITIVE LOGITS
     every
    0.36
     did
    0.34
     manage
    0.29
     Did
    0.29
     raise
    0.29
    ãģ¾ãģĹãģŁ
    0.28
    åŁºäºİ
    0.28
     move
    0.27
    Every
    0.27
     Every
    0.27
    Activations Density 0.014%

    No Known Activations