Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Qwen3-4B
    3. 1-TRANSCODER-HP
    4. 1
    Prev
    Next
    INDEX
    Explanations

    ws

    np_max-act · gemini-2.0-flash

    tokens containing the letter “w” (lowercase or uppercase), especially within common letter clusters like ew/ow/we and ...ews substrings.

    oai_token-act-pair · gpt-5Triggered by @liuliujiujiu05
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_1.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.1.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    庶
    -0.27
    èĴ¯
    -0.26
     provision
    -0.25
    ä¸įè¿ľ
    -0.24
    级以ä¸Ĭ
    -0.24
    ouch
    -0.24
    è¡·
    -0.24
     yö
    -0.24
     goal
    -0.23
    rst
    -0.23
    POSITIVE LOGITS
    æį¢ä¸ª
    0.29
    zeich
    0.27
    haul
    0.25
    åĬĽåѦ
    0.25
    epam
    0.25
    ogi
    0.24
    第ä¸ĢæĿ¡
    0.24
    tout
    0.24
    æ±¶
    0.24
    prites
    0.24
    Activations Density 20.334%

    No Known Activations