© 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. 31-TRANSCODER-HP
    7. 18241
    Prev
    Next
    INDEX
    Explanations

    Mixed text excerpts

    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
    Not in Any Lists

    No Comments

    Negative Logits
     Required
    -0.30
     required
    -0.28
    ulent
    -0.27
    elter
    -0.27
    required
    -0.26
    REQUIRED
    -0.26
    Fault
    -0.25
     Fault
    -0.24
    ulence
    -0.24
    ç§ĺ书
    -0.24
    POSITIVE LOGITS
    .skip
    0.29
    angen
    0.29
     overlooked
    0.28
     left
    0.28
    avigation
    0.26
     SKIP
    0.26
    ToLeft
    0.26
    åıĺå¾ĹæĽ´åĬł
    0.26
    aktu
    0.25
     term
    0.25
    Activations Density 1.300%

    No Known Activations