© 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. 19458
    Prev
    Next
    INDEX
    Explanations

    Code and documentation snippets

    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
    è¦ģç´§
    -0.26
    aru
    -0.25
    _pat
    -0.25
    arda
    -0.24
     Kapoor
    -0.24
     Pat
    -0.24
    æ±Ŀ
    -0.24
    oire
    -0.24
    Reddit
    -0.24
     PAT
    -0.24
    POSITIVE LOGITS
    课åłĤ
    0.28
     checkBox
    0.28
     plan
    0.28
     gerne
    0.27
    äºĶæĺŁ
    0.27
    esub
    0.27
    helper
    0.26
    apter
    0.26
    å·¥åķĨ
    0.26
    hr
    0.24
    Activations Density 0.002%

    No Known Activations