© 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. 25-TRANSCODER-HP
    7. 108674
    Prev
    Next
    INDEX
    Explanations

    code/documents

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_25.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.25.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     example
    -0.27
     _$
    -0.26
    æĢªçī©
    -0.25
    .getRight
    -0.25
    æŃ¹
    -0.24
    åĢ¡å¯¼
    -0.24
    åIJįåįķ
    -0.24
    绺
    -0.23
     combo
    -0.23
     illustr
    -0.23
    POSITIVE LOGITS
    uhn
    0.29
    æĬķåħ¥
    0.26
    ughters
    0.26
    大夫
    0.26
     '..',
    0.26
    Plate
    0.25
    zero
    0.25
    gin
    0.25
    chrift
    0.24
     Uploaded
    0.24
    Activations Density 0.884%

    No Known Activations