© 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. 129646
    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
    ello
    -0.28
    elog
    -0.26
    inis
    -0.26
    INET
    -0.25
    æķĽ
    -0.25
    ordable
    -0.24
     dum
    -0.24
    è¡ĮåĪĹ
    -0.24
     {:?}",
    -0.24
     compelled
    -0.24
    POSITIVE LOGITS
    说åĪ°è¿ĻéĩĮ
    0.25
    ä¸Ĭå¸Ĥ
    0.25
    çľī头
    0.24
    quer
    0.24
     ideal
    0.24
    çļĦçľ¼ç¥ŀ
    0.24
     Chess
    0.23
    upid
    0.23
    things
    0.23
    åĪĨæŀIJå¸Ī
    0.23
    Activations Density 0.010%

    No Known Activations

    This feature has no known activations.