© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 471
    Prev
    Next
    INDEX
    Explanations

    of [events, cities, prevalence, Jesus]

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

    The neuron activates on numeric tokens—digits and decimal numbers—wherever counts, measurements, or statistics appear.

    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
    Predic
    -0.87
     Predic
    -0.85
     ketika
    -0.85
     impra
    -0.84
    MainPage
    -0.82
    न्
    -0.79
    USI
    -0.79
    ARANCE
    -0.79
    PROOF
    -0.78
    stof
    -0.77
    POSITIVE LOGITS
     supposed
    1.02
     zust
    1.01
     about
    0.95
     ending
    0.95
     its
    0.94
     story
    0.86
     facts
    0.84
    icorn
    0.83
    եւ
    0.83
     culmination
    0.82
    Activations Density 0.106%

    No Known Activations