© 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. 19396
    Prev
    Next
    INDEX
    Explanations

    hand

    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
    ç»Ļ
    -0.51
    給
    -0.47
    ç»ĻäºĨ
    -0.38
    ç»ĻåŃ©åŃIJ
    -0.38
     gave
    -0.36
    ç»Ļå®Ŀå®Ŀ
    -0.36
    ç»Ļ她
    -0.35
    給äºĪ
    -0.34
    ç»ĻäºĪ
    -0.33
    ç»ĻäºĪäºĨ
    -0.33
    POSITIVE LOGITS
    ite
    0.28
     Hel
    0.28
    acked
    0.26
    áŀĤ
    0.25
     adapt
    0.25
    ieren
    0.25
    æĮijåīĶ
    0.25
    IED
    0.25
    bags
    0.24
    //===
    0.24
    Activations Density 0.799%

    No Known Activations