© 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. 20-TRANSCODER-HP
    7. 141258
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     finish
    -0.28
     setC
    -0.27
    CategoryId
    -0.27
    ä¸Ģçīĩ
    -0.27
    æŶåĪ»
    -0.25
    ä¸ĢåĪĨ
    -0.25
     reap
    -0.25
    categoryId
    -0.24
    帽
    -0.24
    俾
    -0.24
    POSITIVE LOGITS
    eko
    0.32
    eka
    0.27
    lete
    0.26
    ç£ħ
    0.26
    å§ĵ
    0.26
    å¯Ĩçłģ
    0.26
     dynam
    0.24
    (rb
    0.24
    doll
    0.23
    è¿ĩå¤ļ
    0.23
    Activations Density 8.421%

    No Known Activations

    This feature has no known activations.