© 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. 116715
    Prev
    Next
    INDEX
    Explanations

    trains and stations

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

    The neuron fires strongly whenever it sees the word “train” (including in phrases like “train wreck” or “train of action potentials”), i.e. it detects mentions of “train.”

    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
     waarschijnlijk
    -0.84
    ettbewerb
    -0.82
    늙
    -0.81
    ipro
    -0.80
    donde
    -0.80
    obrázek
    -0.79
    渃
    -0.79
    zdjęcie
    -0.79
     pabrik
    -0.79
     moeilijk
    -0.78
    POSITIVE LOGITS
     train
    1.70
     Train
    1.52
     trains
    1.32
    Train
    1.25
    TRAIN
    1.20
    列車
    1.20
     TRAIN
    1.20
    wreck
    1.13
     wreck
    1.11
     Trains
    1.10
    Activations Density 0.013%

    No Known Activations