© 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. 18-TRANSCODER-HP
    7. 89677
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_18.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.18.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    åıĤ
    -0.27
    åij¼ãģ°
    -0.26
     formats
    -0.25
    ooks
    -0.25
    }s
    -0.25
    ipo
    -0.25
    _MIC
    -0.24
    =headers
    -0.24
    (Call
    -0.23
    åĽŁ
    -0.23
    POSITIVE LOGITS
    ä¸įåĪ°
    0.28
    -pad
    0.28
     defaultMessage
    0.27
     dig
    0.26
    lessness
    0.26
    éĻĢ
    0.25
    goto
    0.25
     (~(
    0.25
    overs
    0.25
    让æĪij
    0.25
    Activations Density 14.816%

    No Known Activations

    This feature has no known activations.