© 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. 101702
    Prev
    Next
    INDEX
    Explanations

    "Occ" related suffixes

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

    The neuron strongly activates on the three‐letter subword “occ” (as in occult, occasional, occurrence) whenever it appears as its own token.

    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
    鶘
    -3.30
     \%)
    -3.22
    婞
    -3.20
    ᯾
    -3.13
    ى
    -2.97
    guitar
    -2.88
    圌
    -2.86
    ！！！”
    -2.84
    媗
    -2.81
    
    -2.80
    POSITIVE LOGITS
     
    4.66
     '
    4.00
     This
    3.31
     When
    3.22
     That
    3.09
     scrutin
    3.06
            
    3.02
    h
    2.83
     During
    2.78
     Alongside
    2.75
    Activations Density 0.009%

    No Known Activations