© 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. 32-TRANSCODER-HP
    7. 42535
    Prev
    Next
    INDEX
    Explanations

    scientific study

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    _eng
    -0.27
    å¼ĤæĢ§
    -0.27
    argar
    -0.25
    bugs
    -0.25
     leven
    -0.24
     strangers
    -0.24
    -eng
    -0.24
    é¼łæłĩ
    -0.24
     Stam
    -0.24
    servername
    -0.24
    POSITIVE LOGITS
    raising
    0.28
    æ´Ĺ
    0.27
    头
    0.26
    raise
    0.25
    depth
    0.24
     dio
    0.24
    ZE
    0.24
    æµģ
    0.24
    _pause
    0.24
    éĴ²
    0.24
    Activations Density 0.002%

    No Known Activations