© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 111157
    Prev
    Next
    INDEX
    Explanations

    digits followed by URL characters

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

    percent-encoded characters in URLs, particularly the digit '2' that commonly appears in URL encoding sequences (like %20, %22, %27, %40).

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

    The neuron strongly fires on hexadecimal characters (0–9 and A–F) used in percent‐encoded strings.

    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
    -2.53
    If
    -2.50
     It
    -2.33
    2
    -2.30
    5
    -2.23
    There
    -2.22
    Another
    -2.16
    3
    -2.13
    After
    -2.11
    </h3>
    -2.09
    POSITIVE LOGITS
    
    2.31
    圌
    2.23
    犼
    2.16
    跄
    1.99
    眛
    1.96
     molti
    1.93
     ,‬
    1.91
    makeat
    1.91
    
    1.90
     lã
    1.88
    Activations Density 0.013%

    No Known Activations