© 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. 22-TRANSCODER-HP
    7. 98332
    Prev
    Next
    INDEX
    Explanations

    Steel

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ä¸įåĩº
    -0.26
    çļĦå¿ĥæĥħ
    -0.25
    usc
    -0.25
     Seymour
    -0.24
    cka
    -0.24
    Own
    -0.24
    人大
    -0.23
    olders
    -0.23
     Loc
    -0.23
    sov
    -0.23
    POSITIVE LOGITS
     Arabia
    0.27
     quadrant
    0.24
    ully
    0.22
    yun
    0.22
     pore
    0.22
    ness
    0.22
    kill
    0.21
    ÙĬØ©
    0.21
    licht
    0.21
    以åıĬåħ¶ä»ĸ
    0.21
    Activations Density 0.136%

    No Known Activations