© 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. 129723
    Prev
    Next
    INDEX
    Explanations

    Code/technical content

    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
    iffin
    -0.29
    kind
    -0.27
    Facade
    -0.26
    ionales
    -0.26
    opoly
    -0.26
    ipop
    -0.26
    GOP
    -0.25
    opcion
    -0.25
    ĵį
    -0.24
    imitive
    -0.24
    POSITIVE LOGITS
    :
    0.28
    .
    0.27
    ÑģоÑĤ
    0.24
     bro
    0.24
    å¸ĥé²ģ
    0.24
    -
    0.24
    ----------------
    0.24
    el
    0.24
     Bro
    0.23
    æķĻè®Ń
    0.23
    Activations Density 0.013%

    No Known Activations