© 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. 19352
    Prev
    Next
    INDEX
    Explanations

    United

    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
    .fre
    -0.32
    eren
    -0.26
    inium
    -0.25
    æ¾¹
    -0.25
    ä»°
    -0.25
    ]%
    -0.25
     `%
    -0.24
    iktig
    -0.23
     COMMIT
    -0.23
    enal
    -0.23
    POSITIVE LOGITS
    ir
    0.28
    og
    0.28
    cene
    0.27
    ÙĪج
    0.27
     sticky
    0.26
    oga
    0.25
     rad
    0.25
     glass
    0.24
    iro
    0.24
     Inform
    0.24
    Activations Density 0.185%

    No Known Activations