© 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. 88568
    Prev
    Next
    INDEX
    Explanations

    Numbers and calculations

    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
    èī¯
    -0.26
    setMessage
    -0.25
    æµĨ
    -0.25
     vi
    -0.25
    itations
    -0.24
    éŨåīį
    -0.24
    stell
    -0.24
    å¥ĩ
    -0.24
    set
    -0.23
    erval
    -0.23
    POSITIVE LOGITS
    ibur
    0.27
    æĢĽ
    0.26
    ðĿIJº
    0.25
    ä¸ĩæĪ·
    0.25
    iers
    0.23
    rowned
    0.23
    æĿŃ
    0.23
    ampionship
    0.23
     Rank
    0.23
    egis
    0.23
    Activations Density 0.009%

    No Known Activations