© 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. 19387
    Prev
    Next
    INDEX
    Explanations

    Academic/technical content

    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
    á¿ĸ
    -0.26
    CLA
    -0.26
    ä½ij
    -0.26
     reli
    -0.25
     Bundy
    -0.25
    (Uri
    -0.24
    (Sprite
    -0.24
     sl
    -0.24
    Uniform
    -0.24
    peria
    -0.24
    POSITIVE LOGITS
    ä¸ĢåŃĹ
    0.30
    ä¸Ģçĵ¶
    0.27
    arden
    0.27
    çĵ¶åŃIJ
    0.26
    onn
    0.25
    fragment
    0.25
    onz
    0.24
    ayah
    0.24
    榨
    0.24
     bản
    0.24
    Activations Density 0.046%

    No Known Activations