© 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. 87723
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    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
     excuse
    -0.27
    kn
    -0.26
    igo
    -0.26
    éħįåIJĪ
    -0.26
    s
    -0.26
    æķ
    -0.25
     raised
    -0.25
    åķ¬
    -0.25
     following
    -0.24
     bounce
    -0.24
    POSITIVE LOGITS
    祷
    0.33
    erview
    0.27
     slut
    0.25
    æľºæ²¹
    0.25
    procs
    0.25
    caller
    0.24
     anál
    0.24
    -syntax
    0.23
    increments
    0.23
    ubern
    0.23
    Activations Density 0.048%

    No Known Activations

    This feature has no known activations.