© 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. 10-GEMMASCOPE-RES-131K
    7. 70878
    Prev
    Next
    INDEX
    Explanations

    yesterday followed by punctuation or event

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

    The neuron fires strongly on the word “Yesterday,” marking mentions of the previous day.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    侂
    -2.70
     tetapi
    -2.70
    鸰
    -2.56
    皤
    -2.55
    namentales
    -2.55
    柺
    -2.39
    儅
    -2.38
    絝
    -2.34
    魮
    -2.31
    翆
    -2.31
    POSITIVE LOGITS
    .
    4.09
     but
    2.81
    ↵
    2.78
    The
    2.50
    It
    2.41
    鹇
    2.38
    "
    2.28
    之人
    2.23
     joking
    2.23
     Then
    2.23
    Activations Density 0.007%

    No Known Activations