© 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. 44605
    Prev
    Next
    INDEX
    Explanations

    explaining reasons or definitions

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

    This neuron detects discourse markers that introduce explanations or definitions (e.g. words like “meaning,” “being,” “motivation,” “point,” etc.).

    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
     was
    -1.94
    ]
    -1.67
     n
    -1.55
     for
    -1.47
     on
    -1.47
     seems
    -1.45
     in
    -1.42
     p
    -1.39
     has
    -1.37
     gritty
    -1.35
    POSITIVE LOGITS
    鹟
    1.64
     髪型
    1.54
    饃
    1.50
    
    1.45
     spektak
    1.44
    tuu
    1.43
     délais
    1.41
     opérateur
    1.41
    瘣
    1.40
     automó
    1.39
    Activations Density 0.026%

    No Known Activations