© 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. 30-TRANSCODER-HP
    7. 96772
    Prev
    Next
    INDEX
    Explanations

    focus

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     concentrating
    -0.31
     concentrates
    -0.30
    focused
    -0.29
    è¶ģçĿĢ
    -0.28
    ä¸ĵ注äºİ
    -0.27
    (flags
    -0.27
    ÑĦи
    -0.26
     focusing
    -0.26
     Concent
    -0.25
    åĽŀäºĭ
    -0.25
    POSITIVE LOGITS
     more
    0.46
     harder
    0.38
    æĽ´å¤ļ
    0.37
    more
    0.35
     More
    0.34
    æĽ´å¤ļçļĦ
    0.33
     hard
    0.32
     better
    0.32
    æĽ´
    0.32
     tougher
    0.31
    Activations Density 0.113%

    No Known Activations