© 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. 19000
    Prev
    Next
    INDEX
    Explanations

    "old"

    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
     decade
    -0.35
    dec
    -0.33
     Dec
    -0.33
    (dec
    -0.32
     tens
    -0.31
    åįģå¹´
    -0.30
     dec
    -0.30
    ONES
    -0.29
    Dec
    -0.29
     decades
    -0.28
    POSITIVE LOGITS
     old
    0.44
    -old
    0.37
    old
    0.36
    èĢģ
    0.34
     oldest
    0.33
     older
    0.32
    	old
    0.32
    (old
    0.31
    older
    0.30
    .old
    0.30
    Activations Density 0.075%

    No Known Activations