© 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. 19511
    Prev
    Next
    INDEX
    Explanations

    carbon dioxide

    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
     wor
    -0.25
     bre
    -0.25
    å·®
    -0.24
    åĻ¬
    -0.24
     Inflate
    -0.24
    度
    -0.23
    èģĮ
    -0.23
     AAC
    -0.23
     confer
    -0.23
    ugg
    -0.23
    POSITIVE LOGITS
     CO
    0.50
    äºĮæ°§åĮĸ碳
    0.49
    CO
    0.49
    _CO
    0.42
     dioxide
    0.39
    carbon
    0.38
     carbon
    0.37
    KO
    0.36
    Co
    0.35
     Carbon
    0.34
    Activations Density 17.724%

    No Known Activations