© 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. 27-TRANSCODER-HP
    7. 57657
    Prev
    Next
    INDEX
    Explanations

    Code and forum posts

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_27.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.27.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     understanding
    -1.43
    çIJĨ解
    -1.37
     understand
    -1.34
     understood
    -1.24
     Understanding
    -1.20
    Understanding
    -1.14
     understands
    -1.13
    çļĦçIJĨ解
    -1.05
    çIJĨ解åĴĮ
    -1.04
     Understand
    -1.00
    POSITIVE LOGITS
    壳
    0.27
    SPA
    0.26
    ROME
    0.26
    аниз
    0.25
    é¹ŀ
    0.25
    гÑĢамм
    0.24
    :black
    0.24
     hitting
    0.24
    éģĩ
    0.24
    TURE
    0.23
    Activations Density 0.000%

    No Known Activations