© 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. 43567
    Prev
    Next
    INDEX
    Explanations

    science and technology

    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
    å¹²
    -0.31
    schemas
    -0.30
    åij¼å£°
    -0.28
    är
    -0.27
    rys
    -0.26
    >Delete
    -0.25
    ARC
    -0.25
    .fasterxml
    -0.25
    è¯Ŀ说
    -0.25
     sem
    -0.24
    POSITIVE LOGITS
    å¤į
    0.28
    dirname
    0.26
    ед
    0.25
    nut
    0.25
     nut
    0.25
    Nut
    0.24
    mite
    0.24
    èĩ³å°Ĭ
    0.24
    墨
    0.24
     Nut
    0.23
    Activations Density 0.004%

    No Known Activations