© 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. 108787
    Prev
    Next
    INDEX
    Explanations

    Technical documents and code snippets

    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
    éĺª
    -0.28
    ibaba
    -0.27
    享
    -0.26
    intValue
    -0.25
    olec
    -0.25
     setSupportActionBar
    -0.25
    èĦIJ
    -0.25
    >[]
    -0.24
     Relief
    -0.24
    骥
    -0.24
    POSITIVE LOGITS
    ElementsBy
    0.28
    erb
    0.27
    ByEmail
    0.25
    геÑĢ
    0.25
    .sb
    0.25
    obbies
    0.25
    สà¸ĩ
    0.25
     arm
    0.25
    quire
    0.25
     ele
    0.24
    Activations Density 0.924%

    No Known Activations