© 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. 25-TRANSCODER-HP
    7. 107205
    Prev
    Next
    INDEX
    Explanations

    results findings

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_25.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.25.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    æŃ£æĺ¯
    -0.32
    ä¹ŁæŃ£æĺ¯
    -0.31
    æĪIJåĬŁçļĦ
    -0.29
    æŃ£æĺ¯åĽłä¸º
    -0.27
    _successful
    -0.27
    ocratic
    -0.27
    羡æħķ
    -0.27
    ueblo
    -0.26
    .constraints
    -0.26
     principals
    -0.26
    POSITIVE LOGITS
    èĪĮ
    0.30
     khỼp
    0.27
    ertz
    0.26
    ì°į
    0.26
     overlap
    0.25
     overd
    0.25
    bie
    0.25
    bez
    0.24
    ä¼Ł
    0.23
     overlaps
    0.23
    Activations Density 0.975%

    No Known Activations