© 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. 34-GEMMASCOPE-RES-131K
    7. 25637
    Prev
    Next
    INDEX
    Explanations

    separators and underscores

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

    The neuron fires on runs of repeated non‐letter characters (especially underscores or dashes, often with spaces) used as ASCII-style horizontal separators or “ruler” lines.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    ={
    -0.95
    zesse
    -0.91
    moldura
    -0.85
    一品
    -0.85
    Referenties
    -0.85
    𒊹
    -0.84
    циони
    -0.84
    ="#">
    -0.84
    }
    -0.84
    그러나
    -0.83
    POSITIVE LOGITS
     _
    1.65
     .
    1.48
     ,
    1.21
     ,,
    1.13
     __
    1.10
         
    1.09
     ...
    1.06
          
    1.02
        
    1.01
       
    1.00
    Activations Density 0.033%

    No Known Activations