© 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. 85489
    Prev
    Next
    INDEX
    Explanations

    capital letters, roman numerals, and names

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

    The neuron responds strongly to any two‐character token, i.e. it picks out exactly two‐letter chunks.

    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
    .
    -3.08
    Both
    -2.72
    Several
    -2.55
    Often
    -2.53
    Some
    -2.48
    Nearly
    -2.48
    Those
    -2.45
    From
    -2.45
    Probably
    -2.42
    Almost
    -2.41
    POSITIVE LOGITS
    '
    2.83
    
    2.44
     to
    2.30
    
    2.28
    
    2.27
    のイメージ
    2.25
     breathtaking
    2.20
     shimmering
    2.19
     teniendo
    2.19
     dando
    2.19
    Activations Density 0.013%

    No Known Activations