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. 23-TRANSCODER-HP
    4. 3072
    Prev
    Next
    INDEX
    Explanations

    hy

    np_max-act · gemini-2.0-flash

    occurrences of the character sequence “hy” (especially as a capitalized prefix or standalone token) within words and names.

    oai_token-act-pair · gpt-5Triggered by @liuliujiujiu05
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_23.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.23.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     portrayed
    -0.29
    pm
    -0.28
    inals
    -0.28
    ingly
    -0.28
     pm
    -0.27
    å¥Ĺ
    -0.26
    ored
    -0.25
    容å¿į
    -0.25
    appen
    -0.24
    ä¸ĢæĬĬ
    -0.24
    POSITIVE LOGITS
    Park
    0.30
    generation
    0.29
    ogene
    0.28
    otron
    0.28
    æĴ¥
    0.27
     можеÑĤе
    0.27
    ogen
    0.26
     Park
    0.25
    adian
    0.25
    ration
    0.24
    Activations Density 0.168%

    No Known Activations