© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 25-TRANSCODER-HP
    7. 107540
    Prev
    Next
    INDEX
    Explanations

    wasn

    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
    Not in Any Lists

    No Comments

    Negative Logits
    满足
    -0.28
    abel
    -0.28
    ä¸įèī¯
    -0.26
      ↵
    -0.25
    æīĩ
    -0.25
     and
    -0.25
    and
    -0.24
    èĥ¡åIJĮ
    -0.24
    .visible
    -0.24
    ï¼Ī
    -0.24
    POSITIVE LOGITS
    sst
    0.28
     ACA
    0.26
    *dt
    0.25
    worm
    0.25
    ERA
    0.25
    ToJson
    0.25
    åŃķæľŁ
    0.25
    ALER
    0.25
     ARR
    0.25
    _allocated
    0.25
    Activations Density 4.364%

    No Known Activations