© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 18-TRANSCODER-HP
    7. 90017
    Prev
    Next
    INDEX
    Explanations

    Sentence beginnings/endings, punctuation

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_18.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.18.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æĶ¾å®½
    -0.31
     freedom
    -0.30
    eh
    -0.29
    amine
    -0.29
     Liberation
    -0.29
     non
    -0.27
    fre
    -0.27
     liberty
    -0.27
    ä¸įåıĹ
    -0.25
    çĥĥ
    -0.25
    POSITIVE LOGITS
     zar
    0.29
    -exclusive
    0.27
    quito
    0.26
     ritual
    0.26
    pictures
    0.26
     collects
    0.25
    BOSE
    0.25
    acted
    0.25
    styl
    0.25
    (IDC
    0.25
    Activations Density 15.723%

    No Known Activations