© 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. 33-TRANSCODER-HP
    7. 53785
    Prev
    Next
    INDEX
    Explanations

    war

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    —are
    -0.24
    exels
    -0.24
    æĪĺ
    -0.23
    onz
    -0.23
    asket
    -0.23
     shift
    -0.23
    Animating
    -0.23
    _REST
    -0.23
    auc
    -0.22
    æĪĺçģ«
    -0.22
    POSITIVE LOGITS
     broke
    0.37
    -breaking
    0.31
    r
    0.31
     lo
    0.30
     r
    0.29
     reverted
    0.29
     reunited
    0.28
    Issue
    0.28
     followed
    0.27
     breaking
    0.27
    Activations Density 0.026%

    No Known Activations