© 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. 33-TRANSCODER-HP
    7. 54155
    Prev
    Next
    INDEX
    Explanations

    violence and weapons

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    å¼¹
    -0.53
    å½Ī
    -0.38
     magazine
    -0.38
     clip
    -0.32
     magazines
    -0.31
    æ®ī
    -0.30
     bullet
    -0.30
    å¼
    -0.30
    Pow
    -0.29
     jam
    -0.28
    POSITIVE LOGITS
     arms
    0.29
    Forge
    0.27
    аÑĢÑĮ
    0.26
    :@
    0.25
    åIJı
    0.25
    <dd
    0.25
    ignKey
    0.24
     arm
    0.24
    æ»ļ
    0.24
    éİ®
    0.24
    Activations Density 0.125%

    No Known Activations