© 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. 107188
    Prev
    Next
    INDEX
    Explanations

    late

    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
    å¥Ħ
    -0.25
     Bris
    -0.25
    åĢĴ
    -0.24
    ulus
    -0.24
     dress
    -0.24
     indirectly
    -0.24
     prepares
    -0.24
     hack
    -0.23
    -&
    -0.23
     opaque
    -0.23
    POSITIVE LOGITS
     hardest
    0.31
    _hard
    0.27
    ritte
    0.26
    硬
    0.26
    adir
    0.25
    irs
    0.24
    LOTS
    0.24
    [Unit
    0.24
    æĪIJ交
    0.24
     deepest
    0.24
    Activations Density 4.845%

    No Known Activations