© 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. 31-TRANSCODER-HP
    7. 19103
    Prev
    Next
    INDEX
    Explanations

    War and oppressive regimes

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æį£
    -0.29
     dign
    -0.28
    irty
    -0.27
    моÑĢ
    -0.27
    绢
    -0.27
     isol
    -0.26
    .createComponent
    -0.25
    urrencies
    -0.24
    erase
    -0.24
    ylum
    -0.24
    POSITIVE LOGITS
    çĶ¥
    0.30
    [random
    0.27
    éĴĪ对æĢ§
    0.26
    ãĥIJãĥ«
    0.25
    åħ¶ä¸Ńä¹ĭä¸Ģ
    0.25
    ÑģÑĥÑĤ
    0.25
    bet
    0.25
    so
    0.25
    ç©°
    0.25
     Bet
    0.25
    Activations Density 0.494%

    No Known Activations