© 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. 18369
    Prev
    Next
    INDEX
    Explanations

    base

    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
     indoors
    -0.33
    室åĨħ
    -0.31
    IDL
    -0.29
    æ¾¹
    -0.28
     inland
    -0.26
    prompt
    -0.25
    åijĭ
    -0.24
     Collider
    -0.24
    èĥľä»»
    -0.24
    oshi
    -0.24
    POSITIVE LOGITS
    .base
    0.32
     base
    0.32
    base
    0.31
     Base
    0.30
     Roman
    0.29
    åŁºåľ°
    0.29
     bases
    0.29
    æīĵãģ¡
    0.29
    _base
    0.29
    ç½®
    0.29
    Activations Density 0.195%

    No Known Activations