© 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. 25632
    Prev
    Next
    INDEX
    Explanations

    never seen before

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

    The neuron detects adverbs and related words that emphasize rarity or novelty—terms like “never,” “rarely,” “ever,” and “before.”

    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
     cuáles
    -0.87
    pă
    -0.80
     télécharge
    -0.79
    痨
    -0.79
    manned
    -0.78
    tiva
    -0.77
     компьютера
    -0.76
     redak
    -0.76
     kön
    -0.75
    ኖ
    -0.75
    POSITIVE LOGITS
     seen
    3.50
    seen
    2.64
    Seen
    2.22
     Seen
    2.14
     visto
    1.78
     previously
    1.77
     SEEN
    1.77
    见过
    1.73
     before
    1.72
     vistos
    1.68
    Activations Density 0.037%

    No Known Activations