© 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. 108227
    Prev
    Next
    INDEX
    Explanations

    code fragments

    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
    gap
    -0.30
     Recommend
    -0.27
    ackers
    -0.26
     Recomm
    -0.25
    DONE
    -0.25
     Plates
    -0.25
    erval
    -0.25
    CharCode
    -0.24
     Mellon
    -0.24
    èī²
    -0.24
    POSITIVE LOGITS
    mus
    0.33
    *ft
    0.31
     pressure
    0.29
    åİĭåĬĽ
    0.27
    æ¯ıæĹ¥
    0.26
     hanging
    0.26
     media
    0.25
    å®ĥçļĦ
    0.25
    pressure
    0.25
    etic
    0.25
    Activations Density 0.058%

    No Known Activations