© 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. 19456
    Prev
    Next
    INDEX
    Explanations

    improving, advancing

    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
    åį°èĬ±
    -0.28
     taxation
    -0.27
    æľĿ
    -0.27
    è¬Ŀ
    -0.26
    spd
    -0.26
    天é¹ħ
    -0.26
     opioids
    -0.25
    iete
    -0.25
    ÑĤок
    -0.25
    ç½®
    -0.25
    POSITIVE LOGITS
     integrating
    0.34
     turning
    0.32
     developing
    0.32
     expanding
    0.31
     advancing
    0.30
     converting
    0.29
     combining
    0.28
     advances
    0.28
     adapting
    0.27
     deploying
    0.27
    Activations Density 1.206%

    No Known Activations