© 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. 30-TRANSCODER-HP
    7. 95918
    Prev
    Next
    INDEX
    Explanations

    code and copyright notices

    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
     actual
    -0.29
    红楼
    -0.27
    ä¸ĢåłĨ
    -0.27
     Guill
    -0.25
    å®īåħ¨ç®¡çIJĨ
    -0.25
     positioned
    -0.24
     hammered
    -0.24
    ä¹ĸä¹ĸ
    -0.24
    actual
    -0.24
    (actual
    -0.24
    POSITIVE LOGITS
    ç¨Ģ
    0.30
     dips
    0.29
     dipping
    0.28
     rar
    0.27
    lich
    0.26
     manifested
    0.25
    pbs
    0.25
    éľĩ
    0.24
    èĩŃ
    0.24
    (conf
    0.24
    Activations Density 0.098%

    No Known Activations