© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 129283
    Prev
    Next
    INDEX
    Explanations

    place names ending in 'court' or 'land'

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

    The neuron is picking up non-English “syllable” tokens characteristic of romanized Indian names and words (e.g. ar, un, ka, hi, Mai, ars).

    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
    或
    -1.76
    Predecesor
    -1.73
    гій
    -1.70
     überlegen
    -1.68
     We
    -1.66
    fastjson
    -1.65
    回戦
    -1.62
     Notwithstanding
    -1.61
    issionais
    -1.61
     of
    -1.55
    POSITIVE LOGITS
    eeeeee
    1.77
    comings
    1.70
    himself
    1.66
    thig
    1.65
    sssss
    1.64
    tobe
    1.63
    uuuuu
    1.63
    漚
    1.61
     alpes
    1.61
     mortos
    1.61
    Activations Density 0.004%

    No Known Activations