© 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. 19890
    Prev
    Next
    INDEX
    Explanations

    grade

    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
    unction
    -0.29
     pouring
    -0.28
     Transit
    -0.28
     Means
    -0.27
    afia
    -0.26
    Means
    -0.26
     statuses
    -0.26
     Line
    -0.26
    means
    -0.25
    çĵ¶
    -0.25
    POSITIVE LOGITS
     grade
    0.34
    igar
    0.30
    grade
    0.27
    -grade
    0.27
    grading
    0.27
    roscope
    0.26
    bra
    0.26
    rey
    0.26
     rated
    0.26
     dollar
    0.26
    Activations Density 0.012%

    No Known Activations