© 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. 102487
    Prev
    Next
    INDEX
    Explanations

    polite Japanese requests or conditional clauses

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

    This neuron detects Japanese verb endings and auxiliary forms (e.g. potential, causative, volitional, and polite constructions) attached to verbs.

    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
    娑
    -1.00
     poder
    -0.98
    ológico
    -0.97
    ⧠
    -0.97
     ו
    -0.97
    omers
    -0.95
     observaciones
    -0.95
    ceptions
    -0.91
    execSQL
    -0.91
     gerekiyor
    -0.90
    POSITIVE LOGITS
    のが
    0.89
    如下
    0.87
    
    0.85
     größe
    0.82
     Ugh
    0.82
    Breaking
    0.81
     Décembre
    0.79
     zoals
    0.78
    Dropped
    0.78
    DRO
    0.77
    Activations Density 0.002%

    No Known Activations