© 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. 35-TRANSCODER-HP
    7. 40104
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_35.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.35.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     exc
    -0.29
    èĥĮ
    -0.28
    åĽ½ç±į
    -0.27
    ↵↵↵
    -0.27
    Exc
    -0.27
     Beat
    -0.25
     block
    -0.25
    触æij¸
    -0.25
    订
    -0.25
    ä»°
    -0.24
    POSITIVE LOGITS
    thur
    0.28
    éĿĴ
    0.27
    -display
    0.27
    isoft
    0.27
     giveaway
    0.27
    ertoire
    0.26
    uttle
    0.26
    ettle
    0.25
     originals
    0.25
    泡泡
    0.25
    Activations Density 37.776%

    No Known Activations

    This feature has no known activations.