© 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. 142075
    Prev
    Next
    INDEX
    Explanations

    Months

    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
    apot
    -0.29
     occupation
    -0.28
    볨
    -0.28
    åįłçĶ¨
    -0.27
    typing
    -0.26
     ocup
    -0.25
    è½®
    -0.25
    oles
    -0.25
     Inner
    -0.25
     Shell
    -0.25
    POSITIVE LOGITS
     tats
    0.27
     bubb
    0.26
    Than
    0.25
    ~=
    0.25
    人æŃ»äº¡
    0.24
     airs
    0.23
     staÅĤ
    0.23
    comb
    0.23
    çªĴ
    0.23
     animator
    0.23
    Activations Density 1.453%

    No Known Activations