© 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. 27-TRANSCODER-HP
    7. 58025
    Prev
    Next
    INDEX
    Explanations

    malicious activity

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_27.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.27.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ayed
    -0.29
    æĵĤ
    -0.28
     bred
    -0.27
    çĥŃ度
    -0.27
    屿
    -0.27
     WEIGHT
    -0.26
     ÑģоÑģÑĤ
    -0.26
     bells
    -0.26
    åºĶä»ĺ
    -0.26
    -states
    -0.26
    POSITIVE LOGITS
     steal
    0.30
    åĪłéĻ¤
    0.29
    ä¸Ńæ¯Ĵ
    0.28
    è°IJ
    0.28
     encrypt
    0.28
    车祸
    0.27
    çļĦçľ¼çĿĽ
    0.26
     potentially
    0.26
     delete
    0.26
    åĪł
    0.26
    Activations Density 0.002%

    No Known Activations