© 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. 22-TRANSCODER-HP
    7. 110973
    Prev
    Next
    INDEX
    Explanations

    Technical/Formal language

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    loid
    -0.28
    éŨ
    -0.28
    enberg
    -0.28
     :↵↵↵↵
    -0.26
    ä»Ģä¹Īåij¢
    -0.26
    -inf
    -0.25
    geries
    -0.24
    logue
    -0.24
    ppard
    -0.24
    ombo
    -0.24
    POSITIVE LOGITS
    ãĢĤ
    0.38
    ;
    0.35
    ãĢĤ↵
    0.34
    ।
    0.33
    .↵
    0.30
    ï¼Į并
    0.30
    áŁĶ
    0.29
    ï¼Ľ
    0.29
    ãĢĤ(
    0.29
    )↵
    0.28
    Activations Density 0.011%

    No Known Activations