© 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. 63221
    Prev
    Next
    INDEX
    Explanations

    code snippets and documentation

    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
    åį±
    -0.30
    缸éģĩ
    -0.26
    istry
    -0.26
    çľĭåĪ°
    -0.26
     directs
    -0.25
     interrupted
    -0.25
     metre
    -0.24
    scri
    -0.24
    ration
    -0.24
    guide
    -0.24
    POSITIVE LOGITS
    andan
    0.28
     finalized
    0.27
    .poll
    0.26
    æ¸ħçIJĨ
    0.26
    .flush
    0.26
    åºĶåıĬæŶ
    0.26
    ä¸Ńåħ³
    0.26
    æĢ»ç»ĵ
    0.26
     finalize
    0.25
    _finalize
    0.25
    Activations Density 34.632%

    No Known Activations