© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 3852
    Prev
    Next
    INDEX
    Explanations

    academic subject headings or introductions

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

    The neuron is detecting LaTeX horizontal‐spacing commands that specify a fixed measurement (e.g. “\hspace{7 mm}”).

    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
     because
    -3.16
     undeniably
    -2.97
     While
    -2.89
     exquisitely
    -2.84
     longstanding
    -2.80
     cleverly
    -2.80
     blatantly
    -2.80
     seems
    -2.72
     ostensibly
    -2.70
     genuinely
    -2.69
    POSITIVE LOGITS
     ridiculously
    3.22
     overarching
    3.09
     turmoil
    2.97
     insanely
    2.95
     mediocre
    2.91
    
    2.91
     crucial
    2.89
     masterful
    2.83
     unbelievably
    2.83
     delightfully
    2.81
    Activations Density 0.000%

    No Known Activations