© 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. 55639
    Prev
    Next
    INDEX
    Explanations

    Latvian names and phrases

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

    The neuron strongly activates on words containing extended Latvian characters (macrons) or other non-ASCII letters, effectively detecting Latvian text fragments.

    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
     pazi
    -2.08
     laim
    -1.70
     savu
    -1.62
     darbu
    -1.59
     būt
    -1.56
     darb
    -1.41
     šo
    -1.30
     veik
    -1.25
     burbu
    -1.24
     meil
    -1.24
    POSITIVE LOGITS
     približ
    1.34
     ljudi
    1.27
     potreb
    1.26
     pokra
    1.25
     žele
    1.18
     različ
    1.17
     objav
    1.17
     postav
    1.16
     sveta
    1.15
     žena
    1.14
    Activations Density 0.009%

    No Known Activations