© 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. 19928
    Prev
    Next
    INDEX
    Explanations

    tools

    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
     Sender
    -0.29
    æħĪ
    -0.26
    æĺ¯çľŁçļĦ
    -0.26
    å®ŀåľ¨
    -0.25
     máquina
    -0.25
    ãģĭãģĭãĤĬ
    -0.25
    棹
    -0.25
     machine
    -0.24
     ple
    -0.24
    .Cascade
    -0.24
    POSITIVE LOGITS
    eries
    0.28
    FR
    0.28
    çľ¨çľ¼
    0.26
     vere
    0.26
     pik
    0.26
    ATA
    0.25
    еÑĢи
    0.25
    磶
    0.25
    erie
    0.24
     unfavor
    0.24
    Activations Density 0.017%

    No Known Activations