© 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. 30-TRANSCODER-HP
    7. 96578
    Prev
    Next
    INDEX
    Explanations

    coerc

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    esk
    -0.31
     PR
    -0.28
    =========
    -0.26
     trial
    -0.26
     explicit
    -0.25
     Bor
    -0.25
     pi
    -0.25
     Jarvis
    -0.25
     pr
    -0.25
    .CR
    -0.24
    POSITIVE LOGITS
    ç»ıèIJ¥èĢħ
    0.29
    erce
    0.29
    æĬĹ
    0.27
    Indented
    0.27
    cx
    0.27
    itag
    0.26
    publisher
    0.25
     skÅĤad
    0.25
    .Module
    0.24
    _inc
    0.24
    Activations Density 0.005%

    No Known Activations