© 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. 9545
    Prev
    Next
    INDEX
    Explanations

    AP and locations

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

    The neuron is selectively detecting standalone numeric tokens and uppercase newswire metadata abbreviations (e.g. “AP,” “U.S.,” dateline acronyms, years, scores) rather than normal running text.

    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
     mogli
    -0.98
     dagen
    -0.96
    onymy
    -0.96
     konser
    -0.96
     maintenant
    -0.93
    oriasis
    -0.93
     teraz
    -0.91
    безпе
    -0.91
     manches
    -0.91
     مستوى
    -0.90
    POSITIVE LOGITS
     started
    1.13
     launched
    1.09
     immediately
    1.05
     stayed
    1.05
     asked
    1.05
     initially
    1.01
     erklär
    0.96
     appeared
    0.95
    olkien
    0.95
     anxious
    0.94
    Activations Density 0.007%

    No Known Activations