© 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. 31-TRANSCODER-HP
    7. 19853
    Prev
    Next
    INDEX
    Explanations

    describing current events

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     lost
    -0.28
     IMapper
    -0.27
     Loss
    -0.27
    usty
    -0.26
    loss
    -0.26
     loses
    -0.26
    -loss
    -0.25
    lost
    -0.25
    linky
    -0.24
    -scenes
    -0.24
    POSITIVE LOGITS
     being
    0.43
    being
    0.36
     facing
    0.36
     becoming
    0.34
     repeating
    0.33
     dying
    0.32
     falling
    0.31
     бÑĭÑĤÑĮ
    0.30
     complete
    0.30
     slide
    0.30
    Activations Density 1.136%

    No Known Activations