© 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. 60625
    Prev
    Next
    INDEX
    Explanations

    events or signals

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

    The neuron is responding most strongly to very common tokens—i.e. high‐frequency stop words (like “the,” “to,” etc.) and punctuation/operators (e.g. “.”, “->” in code).

    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
    以下是
    -0.84
    Доб
    -0.77
     diretta
    -0.75
    Expanding
    -0.75
     Dafür
    -0.75
    Сергей
    -0.75
    查詢
    -0.75
    Expanded
    -0.72
     enjoined
    -0.72
    じゃん
    -0.71
    POSITIVE LOGITS
     signaled
    1.00
     events
    0.92
     signaling
    0.90
     Events
    0.88
     WAIT
    0.82
    APC
    0.81
     signals
    0.80
    ownership
    0.79
     signal
    0.78
    EventType
    0.77
    Activations Density 0.019%

    No Known Activations