© 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. 18625
    Prev
    Next
    INDEX
    Explanations

    us

    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
    (IF
    -0.28
     burner
    -0.26
     fame
    -0.25
    å½
    -0.24
    leases
    -0.24
    å¾ĹåĪĨ
    -0.24
     past
    -0.24
    èĩªæ²»
    -0.24
    _MI
    -0.24
     masturbation
    -0.24
    POSITIVE LOGITS
    uke
    0.33
     yat
    0.29
    ulis
    0.26
    å±ı
    0.26
    åįģäºĶæĿ¡
    0.26
    uce
    0.25
    ulation
    0.25
     hangs
    0.25
    ophil
    0.25
    ograms
    0.24
    Activations Density 0.787%

    No Known Activations