© 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. 43125
    Prev
    Next
    INDEX
    Explanations

    say single letters

    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
    lef
    -0.28
    abb
    -0.28
    è¾ĸ
    -0.27
    èļ¤
    -0.26
    bose
    -0.26
     fi
    -0.25
     fucking
    -0.25
     fir
    -0.24
    .mesh
    -0.24
    è°¬
    -0.24
    POSITIVE LOGITS
    mdat
    0.38
    nk
    0.36
    mlin
    0.34
    ns
    0.34
    rs
    0.33
    ls
    0.33
     mx
    0.32
    ctest
    0.32
    ngen
    0.32
    mb
    0.31
    Activations Density 0.012%

    No Known Activations