© 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. 18301
    Prev
    Next
    INDEX
    Explanations

    of

    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
     Marshal
    -0.33
    IME
    -0.27
    æ²¹æ¼Ĩ
    -0.26
    hic
    -0.25
    æĬĦ
    -0.25
     Bug
    -0.25
    磨æįŁ
    -0.25
     demean
    -0.24
    çĸ½
    -0.24
    åĬŀ
    -0.24
    POSITIVE LOGITS
    utow
    0.31
    nier
    0.27
    ä¹ĥ
    0.26
    Edition
    0.25
    ä¸ŃåĽ½ä¼ģä¸ļ
    0.25
     activating
    0.25
    ligt
    0.25
    aso
    0.24
    panels
    0.24
    å¸ĥæĭī
    0.24
    Activations Density 1.097%

    No Known Activations