© 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. 129333
    Prev
    Next
    INDEX
    Explanations

    pattern

    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
    quez
    -0.29
    undry
    -0.26
    zes
    -0.25
    éķĿ
    -0.25
    urers
    -0.24
    ********************************************************************************
    -0.24
    riger
    -0.24
     czy
    -0.23
     cmdline
    -0.23
    ITU
    -0.23
    POSITIVE LOGITS
     unique
    0.31
     distinct
    0.27
     uniforms
    0.27
    ed
    0.26
    çĭ¬çī¹çļĦ
    0.25
    lane
    0.25
     uniqueness
    0.25
    iard
    0.24
    åĶ¯ä¸Ģ
    0.24
    ainless
    0.24
    Activations Density 0.349%

    No Known Activations