© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 31-TRANSCODER-HP
    7. 17750
    Prev
    Next
    INDEX
    Explanations

    states of being

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     right
    -0.47
    right
    -0.44
     about
    -0.38
    åħ³äºİ
    -0.35
     пÑĢав
    -0.34
    about
    -0.34
     Right
    -0.33
    æĽ´å¥½çļĦ
    -0.33
    æĽ´å¤§çļĦ
    -0.33
    RIGHT
    -0.32
    POSITIVE LOGITS
    ully
    0.33
    opus
    0.28
     sharply
    0.26
     strongly
    0.25
    .ipv
    0.25
    å¸ĤåĮº
    0.25
    è®¤çľŁ
    0.25
     patiently
    0.25
     efficiently
    0.24
    çŀ»
    0.24
    Activations Density 0.786%

    No Known Activations