© 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. 18840
    Prev
    Next
    INDEX
    Explanations

    make

    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
    ħ§
    -0.29
    cuts
    -0.27
    DI
    -0.27
    illin
    -0.26
    orth
    -0.26
    ä»»
    -0.25
    }\.[
    -0.25
    зван
    -0.25
    types
    -0.24
    clr
    -0.24
    POSITIVE LOGITS
     Palmer
    0.30
     Gab
    0.29
    å¹³åĿĩæ¯ı
    0.28
    ä¸ĢåĽ¢
    0.28
     Albany
    0.27
     likely
    0.26
     Holy
    0.24
    è±Ĩçĵ£
    0.24
     subdiv
    0.23
    reten
    0.23
    Activations Density 0.148%

    No Known Activations