© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-1.7B
    5. 27-LLAMASCOPE-2-LORSA-16K-K64
    7. 15407
    Prev
    Next
    INDEX
    Explanations

    Method 3 – the top positive logits are terms like Chunk, Cluster, clusters, swarm, Scatter, indicating the neuron fires for grouping/cluster‑related words. **say clustering**

    unknown · unknown
    New Auto-Interp
    Top Features by Cosine Similarity
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    镫
    -18.13
    itore
    -17.00
    _experience
    -17.00
     Hond
    -16.88
    shi
    -16.88
    acist
    -16.63
    Autor
    -16.50
    /ex
    -16.50
    .crt
    -16.25
     experi
    -16.13
    POSITIVE LOGITS
     Chunk
    18.13
     swarm
    17.25
     streaming
    15.88
    線上
    15.50
     cluster
    15.44
     Cluster
    15.25
     CHUNK
    15.06
     clusters
    15.06
     Scatter
    14.88
    -await
    14.75
    Activations Density 0.525%

    No Known Activations