© 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. 34-GEMMASCOPE-RES-131K
    4. 53943
    Prev
    Next
    INDEX
    Explanations

    design and support elements

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

    The neuron strongly activates on multi‐syllable abstract noun forms (nominalizations) ending in common noun‐forming suffixes (–ion, –ment, –ance, –ity, 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_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
     oftentimes
    -0.91
    Rival
    -0.82
     Eurasian
    -0.81
     Надо
    -0.81
     exquisite
    -0.80
     Initialization
    -0.78
     multifaceted
    -0.78
    cerna
    -0.78
     EVERY
    -0.77
    ENGINEERING
    -0.76
    POSITIVE LOGITS
    parece
    1.00
    WARE
    0.96
    recorded
    0.92
     grafik
    0.91
    </sup>
    0.91
     boden
    0.90
    apparently
    0.90
    informed
    0.89
    ASH
    0.88
    <0x8D>
    0.87
    Activations Density 0.001%

    No Known Activations