© 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. 20-TRANSCODER-HP
    7. 107275
    Prev
    Next
    INDEX
    Explanations

    Beginning of sentences/nouns

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    éļ°
    -0.28
    @Table
    -0.27
     Pers
    -0.27
    åĴ©
    -0.26
    亲æĪļ
    -0.25
    isel
    -0.24
    ëĭ¤ê°Ģ
    -0.24
     represent
    -0.24
    åįĸåĩº
    -0.24
     Represent
    -0.23
    POSITIVE LOGITS
    öm
    0.28
    ói
    0.27
    oky
    0.27
    .va
    0.26
     ActionTypes
    0.25
    brains
    0.25
     bourgeois
    0.24
    REW
    0.24
     worth
    0.23
    .struts
    0.23
    Activations Density 28.925%

    No Known Activations