© 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. 32-TRANSCODER-HP
    7. 42689
    Prev
    Next
    INDEX
    Explanations

    chemistry, materials, energy

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    clone
    -0.29
    peat
    -0.29
    reinterpret
    -0.28
    eus
    -0.28
    avail
    -0.26
    èĬľ
    -0.26
    浪费
    -0.25
     Treaty
    -0.25
    ucht
    -0.25
    é«ĺä½İ
    -0.25
    POSITIVE LOGITS
     being
    0.33
    lod
    0.29
     washed
    0.29
    åijĹ
    0.28
    ä¸į失
    0.28
    Being
    0.28
    æĦıåIJij
    0.28
     drawn
    0.27
     carried
    0.27
     knocked
    0.27
    Activations Density 0.002%

    No Known Activations