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. 7-TRANSCODER-HP
    4. 130178
    Prev
    Next
    INDEX
    Explanations

    the

    np_max-act · gemini-2.0-flash

    superlative-degree constructions, particularly phrases indicating something is among the top or best within a category.

    oai_token-act-pair · gpt-5Triggered by @liuliujiujiu05
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_7.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.7.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    _RCC
    -0.28
     Lev
    -0.26
     DR
    -0.25
    æĸľ
    -0.25
    zb
    -0.24
    ght
    -0.24
    fat
    -0.24
    EVER
    -0.24
    ä¸ĢåĪĩéĥ½
    -0.24
     vol
    -0.23
    POSITIVE LOGITS
    ç½ij讯
    0.29
    å¤ĸæı´
    0.28
    å°ı鸣
    0.27
    橹
    0.26
    uhn
    0.26
     lawy
    0.25
    /mail
    0.25
    aron
    0.24
    顺åĪ©
    0.24
    avana
    0.24
    Activations Density 0.001%

    No Known Activations