© 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. 42290
    Prev
    Next
    INDEX
    Explanations

    short

    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
    lit
    -0.27
     lit
    -0.26
    illard
    -0.26
    lijk
    -0.26
    layers
    -0.26
    edly
    -0.26
    .ly
    -0.25
    ä¸ĢåĽŀäºĭ
    -0.25
    TER
    -0.25
    stage
    -0.25
    POSITIVE LOGITS
     fall
    0.38
     fell
    0.37
     falls
    0.36
    falls
    0.34
     falling
    0.33
     FALL
    0.32
     fallen
    0.32
     Fall
    0.30
    fall
    0.28
    æ»ij
    0.28
    Activations Density 0.004%

    No Known Activations