© 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. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 113751
    Prev
    Next
    INDEX
    Explanations

    URLs with path separators

    np_acts-logits-general · gemini-2.5-flash-lite

    forward slashes in URLs and file paths, particularly after domain names or directory segments.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron is spotting URL fragments—especially domain names, top-level domains, and path separators in web addresses.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    
    -1.70
     forbindelse
    -1.53
    führt
    -1.45
     ஒரு
    -1.45
     Администра
    -1.44
     vitales
    -1.43
    ਲ਼
    -1.41
    ༛
    -1.41
    También
    -1.40
     моя
    -1.40
    POSITIVE LOGITS
     dico
    1.50
     competen
    1.50
     imobili
    1.48
    注意事
    1.45
    ementara
    1.42
     elab
    1.41
     of
    1.41
     comunica
    1.39
     compil
    1.38
     niti
    1.37
    Activations Density 0.112%

    No Known Activations