© 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. 22-GEMMASCOPE-RES-131K
    7. 104418
    Prev
    Next
    INDEX
    Explanations

    warning about potential danger

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

    The neuron fires on verbs and modals (and their accompanying conjunctions) that lay out actions, requests, or warnings in procedural or “what-will/happens” statements.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     suspen
    -0.96
     kazak
    -0.93
     marques
    -0.92
     incendi
    -0.91
    thérapie
    -0.91
    粼
    -0.90
    même
    -0.90
    aneamente
    -0.88
     menimbulkan
    -0.88
     تقویت
    -0.87
    POSITIVE LOGITS
     alert
    1.39
     aware
    1.34
     awareness
    1.15
     potential
    1.08
     alerts
    1.06
     timely
    1.06
     potentially
    1.01
     alerted
    0.99
     warn
    0.98
    提醒
    0.96
    Activations Density 0.075%

    No Known Activations