© 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. 108796
    Prev
    Next
    INDEX
    Explanations

    YOU

    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
    elic
    -0.27
    ittel
    -0.26
    rious
    -0.25
     canadian
    -0.25
    thane
    -0.25
    ä¸įä¹ı
    -0.24
     Sey
    -0.24
    thic
    -0.24
    iless
    -0.24
    å¹Ĥ
    -0.24
    POSITIVE LOGITS
     anything
    0.34
     anywhere
    0.32
     any
    0.31
    ando
    0.30
     D
    0.29
     kind
    0.28
    D
    0.28
     make
    0.27
     base
    0.27
     ↵
    0.26
    Activations Density 1.017%

    No Known Activations