© 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. 20-TRANSCODER-HP
    7. 64657
    Prev
    Next
    INDEX
    Explanations

    despite

    np_max-act · gemini-2.0-flash
    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
    Negative Logits
    常è§Ħ
    -0.28
    è°ı
    -0.27
     OnTrigger
    -0.27
     Carn
    -0.26
    ushi
    -0.25
     tess
    -0.25
    ess
    -0.24
    creat
    -0.24
    ingle
    -0.23
    å°º
    -0.23
    POSITIVE LOGITS
     accumulated
    0.27
    acea
    0.26
     accumulating
    0.25
    æ°ijå¿ĥ
    0.25
     Crash
    0.25
     gathered
    0.25
     Vaults
    0.24
    Dream
    0.24
    oord
    0.24
    oler
    0.23
    Activations Density 15.747%

    No Known Activations