© 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. 97057
    Prev
    Next
    INDEX
    Explanations

    safety and security

    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
    Negative Logits
    ئ
    -0.26
     thuê
    -0.24
    ç«ĭåį³
    -0.23
    yms
    -0.23
    Confirmed
    -0.23
    快讯
    -0.23
     immediately
    -0.23
    Pattern
    -0.22
     Principle
    -0.22
    ä¸Ģä½ĵ
    -0.22
    POSITIVE LOGITS
     safety
    0.78
     security
    0.75
    å®īåħ¨
    0.63
     Safety
    0.60
     Security
    0.59
     health
    0.57
    security
    0.56
    ecurity
    0.55
    -security
    0.55
     privacy
    0.54
    Activations Density 0.917%

    No Known Activations