© 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. 19715
    Prev
    Next
    INDEX
    Explanations

    Contract/Ten (variants)

    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
    åĢĻéĢī
    -0.28
    £¼
    -0.25
    IALIZED
    -0.24
    çłĶåĪ¶
    -0.24
    é¢Ĺç²Ĵ
    -0.24
    $core
    -0.24
    èĮİ
    -0.24
    INED
    -0.23
    Recognition
    -0.23
     Candidate
    -0.23
    POSITIVE LOGITS
     TEN
    0.31
     contract
    0.28
     Ten
    0.28
    åįģ
    0.27
     Terra
    0.26
    _ten
    0.26
    emailer
    0.26
     Contract
    0.26
    contract
    0.25
    Tenant
    0.25
    Activations Density 2.372%

    No Known Activations